|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
67.189.207.135
In Reply to: RE: What tubes to use with this Interstage Transformers? posted by megasat16 on August 28, 2016 at 22:08:29
I am unclear as to why this is suddenly about distortion from my signal generator. I am reasonably confident that if I measure the distortion at the 50 ohm output of my signal generator at 20hz into the primary of an output transformer, and then repeat the measurement with a 5K series resistance the distortion levels at the output of the generator will go down.
If for some reason the goal is to keep a constant input level at the DUT then I see where there could be an issue since as AC current increases, more of the source voltage will be lost across the series resistance requiring an ever increasing output from the generator but the entire goal here is to see a realistic frequency response graph. Attempting to see accurate information of a transformer intending to load a 10 by feeding it from a 50 ohm source doesn't give any meaningful information.
dave
Follow Ups:
Dave,
I know you are here with the intention to help. Therefore, I have tried to to reply in my best possible way and asked you to try it yourself. If I reply in any other way, I would be labeled as insulting.
It's so simple that you only need a Function Generator, a Scope, two 10K resistor, a transformer with around 10K ohm Z.
Set your scope to 21V, 20Hz Sine and put a 5K resistor in series with your SigGen and feed it to your 10K transformer.
I intentionally keep quiet on other topics and cascading transformers for a better reason. Bandwidth of DC to Day Light?
If you refuse to build your own proposed circuit, please let's just keep this topics for the search of suitable tubes.
I have no energy to go back and forth over this simple stuff.
Best Regards,
James
.
.
.Thou shall not stand where I type for I carry a bottle of Certified Audiophile Air and a Pure Silver Whip.
OK...
Set my signal generator to 20hz and a 30V peak-peak signal (maximum output)
measured the FFT of just the waveform into the scope and got this:
I then placed the primary of a 10K:16 output transformer in series with the 50R output impedance and put the scope probe across the 16 ohm load resistor. I then inserted a 5K series resistor with the output of the generator and here are the two FFT's
green is the 50R out and Blue is the 5K50 out.
The loss in amplitude is expected from the finite inductance working against the 5K source but for the most part the distortion remained unchanged.
To keep this complete I also ran frequency sweeps from 1hz to 1Khz with both the 50R and 5K50R source impedance and a 16R load and got the following two sweeps.
again as I stated, the higher source impedance readily shows the results of a finite amount of inductance of this transformer which brings us back to my original point that a sweep of a transformer with a 50 ohm source impedance is pointless unless you intend to use it that way in circuit. Your mention of using a 10 which has roughly a 5K source impedance made me point out that it would be wise to repeat the sweeps with a more appropriate test setup.
I still have no idea what you are referring to about increased distortion and how distortion is even relevant to this discussion but I did what you asked and now it is your turn to repeat the frequency sweeps with a 5K series resistor so we can get a realistic idea of how this device really behaves.
you don't have to take a million data points... Just start at 1Khz and then sweep down to the -1dB, -3dB and -6dB points and that will tell a lot.
dave
With the 5K series resistor in place, probe the input Sine Wave and check if it looks anywhere near a sine wave?
.
.
.Thou shall not stand where I type for I carry a bottle of Certified Audiophile Air and a Pure Silver Whip.
It looks exactly like a sine wave. If it didn't the FFT's would certainly show it.
I really wish you could better explain the mechanism that you think will cause all this phantom distortion.
When you do the new frequency sweeps of the LF behavior with the 5K series resistor, I'm sure you can easily document it.
dave
Why aren't you seeing it? It's beyond me. You don't have to do FFT. And your FFT looks way bumpy to me for a sine wave. I don't know why.According to your own circuit with 5K resistor limiting to about 4mA-40mA (actual number depends on the actual source impedance of the FG) of current driving into a parallel circuit that has low value DCR of less than 50-100ohms, unknown value C but a large value of L in series with the rate of charge of 20 cycle, you will have a resulting vector of triangular output due to the charge due to the very limited current. That triangular output is now reflected back to the source and combined into a pretty distorted sine wave.
To overcome this, you either need to increases the frequency (charge the C of the primary winding more rapidly) or increase the current to charge the capacitance of the winding at every cycle. It's exactly what I envisioned and exactly what I see. Or the result is incomplete charge cycle resulting in a triangular shape waveform. It's the same principle as ripple in Power Supply circuit.
What's the point of using such distorted waveform since it'll only make everything worse? If you can suggest me to modify your own circuit, I am sure it'll work and I can try.
.
.
.Thou shall not stand where I type for I carry a bottle of Certified Audiophile Air and a Pure Silver Whip.
Edits: 08/29/16
hey...
Why aren't you seeing it? It's beyond me. You don't have to do FFT. And your FFT looks way bumpy to me for a sine wave.
I don't think I am seeing it because it isn't there. Why cant you post an example rather than a vague description? Plus that FFT ispretty nice for a 20hz sine... the 2nd and third harmonics are approaching -60dB which is in the area of 0.02% distortion.
According to your own circuit with 5K resistor limiting to about 4mA-40mA (actual number depends on the actual source impedance of the FG) of current driving into a parallel circuit that has low value DCR of less than 50-100ohms, unknown value C but a large value of L in series with the rate of charge of 20 cycle, you will have a resulting vector of triangular output due to the charge due to the very limited current. That triangular output is now reflected back to the source and combined into a pretty distorted sine wave.
if this is the case then it should really be easy to see. To make things easier to document I moved to spice to do the circuit and again I couldn't replicate this distortion.
Here are a series of 4 different traces at three different point in the circuit representing a source impedacne of 50 and 5000 ohma and a coil capacitance of 1p, 1n and 1u. If this behavior really exists I would expect ot see at least a hint of it in one of the simulations.
To overcome this, you either need to increases the frequency (charge the C of the primary winding more rapidly) or increase the current to charge the capacitance of the winding at every cycle. It's exactly what I envisioned and exactly what I see.
again please show me some measured or simmed results of this behavior.
Or the result is incomplete charge cycle resulting in a triangular shape waveform. It's the same principle as ripple in Power Supply circuit.
I have yet to see a triangle and pretty thoroughly covered all the things you have said will cause it.
What's the point of using such distorted waveform since it'll only make everything worse?
the waveforrm isn't distorted so no problems with the measured results.
This claim of distortion is yours and I have done everything to try to find it. Rather than sitting there behind a keyboard and insisting it happens why don't you replicate it and post the results? The perfect place to do this would be with the test setup you have already done to measure the transformer in question. Simply measure the output at 1Khz with and without the 5K series resistor and then repeat at 20hz with both. Then a quick pic of the resultant sine for each case should clearly show the problems you insist will happen.
dave
Oh Dave, don't you think you are still confusing the series resistor acts only like voltage dividing / current limiting in the source circuit?
I thought about it last night why yours work and mine doesn't. It is both true for the both of us. It's because DCR of the windings and it's just not fixed on all type of transformers, isn't it?
You are using 10K OPT to measure your plot? Obviously, the DCR is high on them resulting a low load (low current flow through the primary 10K winding) on the source circuit thus creating less voltage drop on your source with 5K series resistor. And less voltage drop from the 5K resistor means more voltage appears at 10K ohm input at the primary. Combining bigger voltage with less current drop in the DCR of primary will result in a complete charge of the internal C in the winding.
Thus, it doesn't reflect the distorted waveform I mentioned earlier and the reason you aren't aware of it before.
I will post pics with all that explained earlier. Please standby.
Regards,
James
.
.
.Thou shall not stand where I type for I carry a bottle of Certified Audiophile Air and a Pure Silver Whip.
What?
The DCR of the windings really doesn't come into play here since it appears in series with the inductance and the inductance dominates at the frequencies in question.
I really cranked up the numbers and may have found what you are referring to. I increased the winding capacitance to 1u (which is never going to happen) and dropped the frequency to 2hz and got this:
the green trace has 1 ohm DCR for the coils and Blue 1K ohm so it is clear to me that the DCR is not the cause. This was for a single wave cycle at startup and after that first cycle it goes away. Below is the first 6 seconds of a transient response.
dave
Dave,DCR is not playing this? Please try your own circuit on the 16R secondary of your OPT to see what I am saying.
You'll find out you have no reason to feeling smart today.
Regards,
James
.
.
.Thou shall not stand where I type for I carry a bottle of Certified Audiophile Air and a Pure Silver Whip.
Edits: 08/30/16
I reversed the transformer in the sims and aside from being severely attenuated the sine still looks good.
I'll await your test results showing this.
dave
You can't sim or model a wrongly assumed SPICE circuit based on the wrong theory and expect to see a correct real world result.
I expect you TRY it with a FG, a scope, real transformers (real stuff)!
The Second pic with the probe before the 5K resistor on the input.The First pic (top) with the probe after the 5K resistor on the input (or the actual input pins).
Top sine (with yellow color) is the actual output from the FG. The bottom sine (with the blue color) is the measured with the probe.
The probe measurement at the 2nd pic (bottom) is before the series 5K resistor. There is a little small voltage drop due to the internal resistance of the hookup wire.
The probe measurement at the 1st pic is after the 5K resistor which shows the sine wave is severely attenuated to the load from the transformer.
More current flowing through the source circuit thus causing a lot of voltage drop across the 5K resistor. And the small voltage with the limited current supply to the transformer creates a severly distorted waveform.
Here is the FFT (real) curve of the 20Hz Sine! Does it look similar to your FFT? I am not familiar with your FFT plot! I need to study what it is showing.
I will pull the Tango X-10S from the box and measure the Primary and Secondary for you a little later.
.
.
.Thou shall not stand where I type for I carry a bottle of Certified Audiophile Air and a Pure Silver Whip.
Edits: 08/30/16 08/30/16 08/30/16
I did measure it in the real world with a scope and a signal genrator. the image you post sure looks like core saturation to me.
I'll look at it again tonight.
dave
Why hit yourself, Dave?
The saturated core needs one of all of these conditions.
1.) DC on the winding.
2.) Turns Short on the winding. (either primary or secondary)
3.) Crazy Low Load Placed on the Secondary at full power it can handle so the core saturates
When did you see such things happen with the pic I posted?
I will post the sweep from the other transformers I promised a little later.
As I said before, don't make me to go over this simple stuff back and forth. I felt like I already used up a lot of energy trying to make you hear what I have been saying all along.
.
.
.Thou shall not stand where I type for I carry a bottle of Certified Audiophile Air and a Pure Silver Whip.
IF this is the type of waveform you are referring to it is coming from saturation. I was able to get this waveform by simply placing 14,000 gauss of AC flux on an 80% nickel core and this waveform is the result of saturation causing non-sinusodal current draw which shows up as a non-sinusodal voltage when drawn through a series resistance. If you increase the value of series resistance the distortion gets worse.
The saturated core needs one of all of these conditions.
do you mean any of these conditions?
1.) DC on the winding.
I would call that excessive DC on the winding. but yes DC can cause saturation.
2.) Turns Short on the winding. (either primary or secondary)
a shorted turn will cause excessive current draw and distortion du to loss of the load but it will not saturate the core. (see my response to 3)
3.) Crazy Low Load Placed on the Secondary at full power it can handle so the core saturates
This is flat out incorrect. Lenz's Law States that AC current through the secondary does not change the flux in a core (which is why a shorted turn will not saturate) both will cause lots of heating and force the inductance to drop to near zero but they will not cause saturation.
The one big cause of saturation you missed was simple AC excitation. Every time you either double the voltage across a coil or cut the excitation frequency in half you double the flux on the core so as you sweep down in frequency you increase the flux and at some point you are going to saturate that puppy which gives rise to exactly the behavior you are describing.
When did you see such things happen with the pic I posted?
I have seen the particular distortion waveform you posted many times and the cause has always been saturation.
Can you tell me the AC flux on the core you were testing when you obtained that waveform?
It was a long day for me yesterday and I have to do a lot today. But I will try to reply to your post when I can glue my back to the chair today.Meanwhile, please also take a closer look and tell me these two waveforms are the same?
They both looks distorted sine but the cause of the distortion and the waveform isn't the same, right?
.
.
.Thou shall not stand where I type for I carry a bottle of Certified Audiophile Air and a Pure Silver Whip.
Edits: 08/31/16
Sure... they may not look identical so here it is with a little lower source Z and a little less AC excitation.
I still strongly suspect the net cause of the distortion in both cases is nonlinear core behavior combined with a source impedance greater than 0 and the cause of the nonlinear core behavior is saturation.
Can you show this type of behavior with an air-cored transformer or inductor?
dave
It would be useful if you show the whole picture (just like I did) instead of just the waveform so I know where you probe at to get that waveform(whether the input or the output or before or after the resistor).No, I don't have any aircore inductor or transformer here to take that measurement. Would the Toroid works?
But may be I can wind DIY aircore inductor myself.
.
.
.Thou shall not stand where I type for I carry a bottle of Certified Audiophile Air and a Pure Silver Whip.
Edits: 08/31/16
Same hookup as you used.
The signal generator has an adjustable output Z so i simply put the 600R output across the the inductor and the scope is also connected across the same two terminals as the signal generator.
If I drop the Zout of the signal generator to 50 ohms the sine looks better. Again this is simply saturation causing distortion of the current waveform which shows up as a distorted voltage waveform when you have any series resistance.
None of this shows up in simulations unless you add a model for the transformer core material into the mix.
dave
I think you are close to a self realization. A little closer than before.
Ask yourself these questions. At what AC voltage (ac current) will my core saturation occurred at 20Hz.
I am sure you know that answer. Try in on a transformer if you aren't sure about an inductor. Try the circuit with a series resistor and without a series resistor.
But this time, hook up (preferably a good Fluke) AC Ammeter in series with your source and check the current drawn into the inductor (or transformer) and observe the waveform on the scope.
By your own definition, the core saturation occurred at higher AC current (or voltage) than the core can take?
Compare your AC current drawn for various tests - one without the series resistor that produces clean waveform and one with a series resistor that have the distorted waveform and let me know your measurements.
I probably won't be able to post tomorrow though.
.
.
.Thou shall not stand where I type for I carry a bottle of Certified Audiophile Air and a Pure Silver Whip.
I think you are close to a self realization. A little closer than before.
If this self-realization is that a saturating transformer causes a distorted voltage waveform through a finite resistance then that is "old hat" and what I would call a concise explanation of what is happening.
The key point here is if 5000 ohms in series with a signal generator causes these issues, then so will a tube with a 5000 ohm RP.
Ask yourself these questions. At what AC voltage (ac current) will my core saturation occurred at 20Hz.
again.... AC current does not saturate a core. A simple proof of this is to take a 10hy inductor operating at an AC flux of 10,000 gauss and increase the gap size to make it a 1HY inductor. You will now draw 10 times the AC current but the flux on the core will remain unchanged.
I am sure you know that answer. Try in on a transformer if you aren't sure about an inductor. Try the circuit with a series resistor and without a series resistor.
it doesn't matter if it is a transformer or an inductor. saturation is saturation.
But this time, hook up (preferably a good Fluke) AC Ammeter in series with your source and check the current drawn into the inductor (or transformer) and observe the waveform on the scope.
By ac ammeter i suspect you mean current probe for viewing the current waveform on a scope.
Compare your AC current drawn for various tests - one without the series resistor that produces clean waveform and one with a series resistor that have the distorted waveform and let me know your measurements.
I guess the 600 pound gorilla in the corner here is why do we care about the measurements of a device well outside its linear operating range? If the 5K series resistor causes gross visible distortion that simply tells you that it is not suited for audio at that particular voltage / frequency combo.
Earlier on in this thread you said that the Tango NC-20 showed more than a 2dB variation 20-20Khz and if measured from a 50 ohm source I suspect either the measurements are flawed or the transformer is defective since with a 1K source it is flat from 10-100Khz.
dave
I have nothing against a real 1K source showing the core saturation. It's what I wanted to see and try it for myself too.I can't, however, take a 1K series resistor, the same as 1K source. It's a limiting resistor and pretty dumb device to call it an AC impedance source.
Check that pic with the Tango curve again for the series resistor; you'll see the AC impedance source is the actual tube and not a series resistor. It's very acceptable to me. This discussion wouldn't never happened if I tried the same circuit with Tango as 1K Rp tube and see a core saturation on the IT I have. It's exactly what I wanted to see.
But whatever reason you think it is, think this yourself. At what voltage the core saturate for a given size transformer. Some small transformer saturate probably at a few volts while others take hundreds of volts to reach core saturation.
If you have a current probe, it's even better. Measure your own circuit with the current probe and post results.
.
.
.Thou shall not stand where I type for I carry a bottle of Certified Audiophile Air and a Pure Silver Whip.
Edits: 09/01/16
For the sake of showing actual low frequency information, a 50 ohms source with a 1K series resistor is no different than a tube with an Rp of 1050 ohms.
A fair bit of good info has been offered up for you to digest here and I have a strong suspicion that there is no way this transformer has any chance whatsoever of loading a 10 / 801 and getting anywhere near flat from 20-20Khz into the grid of any triode. Simply placing a 5K resistor in series with the input signal and 100pf of capacitance across the secondary will quickly tell you if you have any chance of being flat from 20-20Khz with a tube like the 10.
dave
I think the problem lies with the question of what part of the circuit the resistor belongs to when it is placed after the FG (source) but in series with a Load (Transformer).
Dave is seeing that series resistor as part of the source.
I always see that resistor as part of the load circuit and the actual source is the FG.
Can someone simplify it for us once and for all?
.
.
.Thou shall not stand where I type for I carry a bottle of Certified Audiophile Air and a Pure Silver Whip.
If you look at a circuit from two different points you will get two different answers.
from the perspective of the source (ie 50r output of the signal generator) the 5K series resistor becomes part of the load.
from the perspective of the transformer primary, the 5K resistor becomes part of the source impedance.
when you want to analyze a source / load combo you need to take a single point in the circuit and then look at all of the paths to ground. In this case the only thing of interest is the primary of the transformer so looking backwards the source impedance driving it becomes the parallel combo of every path to ground. Take a signal generator with a 50 ohm output and parallel it with a 1 ohm resistor to ground and your source impedance is now just under an ohm.
dave
No matter from where I look at (i.e. From the input pin of the transformer or from the output pin of the FG, I only get the same answer). That resistor is in series with the input pin of the transformer. Thus, creating a current limiter and (variable) voltage drop across it depending on the load impedance of the transformer based on the different frequency.
That's the way I see it no matter where I look at.
Now, what is that wonderful Engineering marvel FG of yours? I would like to look it up and see what kind of circuit it is constructed with.
Have you ever opened it up and see the impedance selector switch selects different resistors or resistance values? And are they the series resistor inside the FG box with the value of 550, 250 and a short wire?
I wonder why no Engineers from HP, and Tektronix never send a box of resistors along with the FG for the end user to transform the source impedance like you mentioned. It would be very useful for sure.
.
.
.Thou shall not stand where I type for I carry a bottle of Certified Audiophile Air and a Pure Silver Whip.
I think you just need to come up with a fundamental understanding of what source impedance is because you keep defining it over and over and then say it cant be.
that pesky variable voltage drop is exactly what you are looking for and is caused by a finite inductance working against a known source impedance. If we drove everything from perfect voltage source then we would get bass response down to DC.
The signal generator is a wavetek 288 and there are switched resistors right at the unbalanced output. I asked a guy at audio precision how they define their source impedance and he also said they switch a "resistor" at the output. I'm sure the guys at HP and tek are also smart enough to put those resistors (or an equivalent sand device) inside the box.
dave
"No matter from where I look at (i.e. From the input pin of the transformer or from the output pin of the FG, I only get the same answer). That resistor is in series with the input pin of the transformer. Thus, creating a current limiter and (variable) voltage drop across it depending on the load impedance of the transformer based on the different frequency."
That's right but what you are not understanding is that the output impedance of a stage is always in series with the "input pin of the transformer".
What this means is, if you have,
1. a source circuit with an output impedance of 5050 ohms
or
2. a source circuit with a output impedance of 50 ohms with a 5000 ohm series resistor added
They will both have a total output impedance of 5050 ohms and will behave the same.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Tre,
What you said is only true if you connect your source to the purely resistive load. You can't even start to think they are a real AC impedance source has an open ended series resistor connected to it.
And how is the load connected to the source (series or parallel)?
Regards,
James
.
.
.Thou shall not stand where I type for I carry a bottle of Certified Audiophile Air and a Pure Silver Whip.
The C and R and L are in parallel with each other and together they constitute the load that's being driven by the source.The load looks like it's in parallel with the 5050 ohm source to me. (the audio generator with 50 ohms of output impedance, in series with the 5000 ohm resistor constitutes the 5050 ohm source)
What does it look like to you?
BTW I ran the sim sweeping 20 to 20kHz with the generator having 5050 ohms output impedance with no series resistor and the way it's shown.
There was no difference in the results.
Either way, a large cap value causes the high frequencies to fall and a low inductance value for the choke causes the low frequencies to fall and all by the same amount at the same frequencies.
It doesn't matter if I start with a generator of high (5050)output impedance with no series resistor or a low output impedance (50) generator with a large value (5000) series resistor. It gives the same result either way.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 09/05/16 09/05/16
I see that as C1, L1, R1 are in parallel with the series resistor R2 of 5K. FG as the AC impedance source.You follow the path the current flows and see what the current sees (i.e branches out).
So, the voltage / current come out of the + end of FG and goes into the 5K resistor in your circuit and then the current see the branch with C1, L1, R1. So, current decided to chop itself off 3 portions and follow through the path according to the value set by R1, Xc, and Xl and then to the Ground.
Regardless of how components are placed (the R2 on top of the FG), where is the actual load begins at? Where is the point the load meets with the FG?
Most sim can't simulate an actual AC impedance source to the actual circuit (most sim software have unlimited current and voltage capability at any impedance you define because of programming limitations and the use of purely mathematical formulae).
If it can simulate a real 5000ohm Zout or a 50ohm Zout, I highly doubt that you will see the same thing between the 5000ohm impedance FG and 50 ohm impedance FG in series with 5K resistor. At the very least, you will see a very varied phase angle between the results from a real 5K ohm FG or source and a 50ohm FG with a 5K series resistors.
And as always, it also depends on where you probe your Scope and what you are looking at.
Regards,
James
.
.
.Thou shall not stand where I type for I carry a bottle of Certified Audiophile Air and a Pure Silver Whip.
Edits: 09/05/16 09/05/16 09/05/16 09/05/16
"If it can simulate a real 5000ohm Zout or a 50ohm Zout,"
It can.
"I highly doubt that you will see the same thing between the 5000ohm impedance FG and 5K ohm impedance FG in series with 5K resistor."
I think you mean a 5000 ohm impedance FG vs. a 50 ohm impedance FG in series with a 5k resistor.
Well, no. The one has a output impedance of 5k ohms and the other has a output impedance of 5.05k ohms.
Other that that they will be the same.
"At the very least, you will see a very varied phase angle between the results from a real 5K ohm FG or source and a 50ohm FG with a 5K series resistors. "
No, the phase (just like the Fc and amplitude loss) was the same as well.
Should I post screen shots of the sims? Or I could post the file and you could run it yourself?
Do you have and use spice? I use LTspiceIV. It's free and pretty easy to use.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Thanks for pointing it out. Yes, I meant 50ohm. It was a typo and I corrected it. I also added a few more lines to that post so please reread it.LTSpice and all of the sim has it's own limitations. First of all, they can presume pretty big numbers (voltage source / current source) since it uses purely mathemtical formula and it has almost unlimited current and voltage capability at any frequency and any impedance you set at.
It's very different from actual voltage source in life.
You need to compare the results with different voltages on the LTSpice model to about 1V to 30V. Compare the results. I don't need to do anything with your Spice sim models. I know where the limitations are at.
I used with play with LTspice and many sims very long time ago. I don't use any of them for a long time so I need to relearn again.
But I can do basics circuit analysis from the top of the head so I don't need to draw up schematic or model in Spice unless it is a very complex circuit I need to deal with.
.
.
.Thou shall not stand where I type for I carry a bottle of Certified Audiophile Air and a Pure Silver Whip.
Edits: 09/05/16
Here's the top one again so you can see them side by side.
The solid line is frequency response and the dashed line is phase.
The probe is at the top of the load(s).
As you can see, they are both the same.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 09/05/16
Tre,Put the scope in time domain mode and compare the measured waveforms. Don't look at bode plotting. It doesn't say much for the reflection from the load to the source.
.
.
.Thou shall not stand where I type for I carry a bottle of Certified Audiophile Air and a Pure Silver Whip.
Edits: 09/05/16
I think what you need to do is build it and measure for yourself.
Dave and I have tried to help you but you just want to argue.
I'm done.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Tre,I did not argue with both of you. I am only correcting the wrong here.
What is the 12V DC source? It is the 12V Battery, right? Where does the voltage come from and how is it measured?
Current comes from the positive terminal of the battery to the negative terminal of the battery, right? And we measured the voltage at the two terminals. No resistor added, right?
What is the 12V AC Source? Same thing, right? We again, don't measure with any series resistor, right?
Is there anywhere in any text book did you see that battery voltage is measured with a series resistor and call that resistor as part of the DC source?
The text book might refer to the battery internal resistance but it's not an actual resistor attached to it.
Such simple principles. Why do I need to argue with you over it?
And you can sim like I said and you can post results if you want. No one is stopping you. But like I said, Sims have it's own limitations.
I can post real test data if you like.
.
.
.Thou shall not stand where I type for I carry a bottle of Certified Audiophile Air and a Pure Silver Whip.
Edits: 09/05/16
"The text book might refer to the battery internal resistance but it's not an actual resistor attached to it."
No it's not.
But a battery with 1 ohm of internal resistance vs. a battery with .5 ohm of internal resistance and a .5 ohm series resistor attached to it will behave the same.
If you have all the answers why did you come here asking questions?
"I can post real test data if you like."
I think you should.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Here are 4 pictures.
The top yellow trace is the actual waveform coming out from the FG. It sets to about 20V at roughly 20Hz.
The bottom blue trace is for the measured waveform with the probe.
Probe hookup / measured points can be seen in the picture.
Take a go at it.
.
.
.Thou shall not stand where I type for I carry a bottle of Certified Audiophile Air and a Pure Silver Whip.
What is the inductance of the IT?BTW In LT spice you can set the resistance of the DC source to any value. Same for a AC source.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 09/05/16
I don't know the inductance at 20Hz on the primary. I also don't know the capacitance at 20Hz on the primary. I also don't know any inductance of the secondary at 20Hz. I also don't know the capacitance of the secondary at 20Hz. Repeat that all the way from for DC to Daylight (whatever it is) Hz.I also don't know the size of the magnetic wire, the turn count of the wire on primary and secondary side, the magnetic core materials (Amophous, Permalloy, HyperMalloy, M9, M6; whatever) used, what core type (EI or C or Toroid) it use; etc.
I know it is 10K Zp transformer with step up ratio of 1:4. I also know the DCR of the primary and secondary but I am not going to tell you yet (not yet now).
Why? Because no transformer manufacturer tells you any inductance value or capacitance value or DCR values.
They only tells you the wattage it can handle, the impedance of the primary and the secondary, the insertion loss (if you are lucky), and the FR and Impedance curve (if you are really lucky).
Almost all of them also failed to mention that maximum voltage it can be operated for and the bandwidth.
So, I can't tell you all of them. But these 4 pics I posted have more than enough information to synthesize what's going on there.
You don't need to churn out numbers for me. Tell me what's going on at the probe point at the input of the transformer from these pics.
Again, it is a 10K primary transformer, open load on the secondary.
Yes, I recall you can set the values. But it doesn't mean it's almost accurately taken into account.
Anyway, I did tell you to try at different voltage levels and observe the waveform like in my screen shot.
.
.
.Thou shall not stand where I type for I carry a bottle of Certified Audiophile Air and a Pure Silver Whip.
Edits: 09/05/16
Internal resistance of the battery is not the same as the external series resistors! Some battery may have 0.0001ohms, some may have a little higher.
But in SIM, battery have extremely low resistance value like 0. A little smarter sim let you define the resistance.
Is it the same for you?
Yes, I will post pics. Do the AC analysis for me when I do.
.
.
.Thou shall not stand where I type for I carry a bottle of Certified Audiophile Air and a Pure Silver Whip.
Other than me disagreeing with your theory of using 5K series resistor, I am just as interested to see how it performs with an actual 5K or 10K source.
Any other method do you have in mind other than building an actual circuit in a short time? I can prototype a circuit but time is not favorable for me right now.
.
.
.Thou shall not stand where I type for I carry a bottle of Certified Audiophile Air and a Pure Silver Whip.
So, you are saying whatever resistor connected to the source becomes the actual source including the resistor value?
50ohms impedance with 1K series is 1050ohms source and 5K becomes 5050ohms source?
So, why don't we just all use 5K ohm OPT in series with a 5K ohm resistor and call it a true 10K ohm transformer? After all, it's much easier to make better performing transformer with lower Primary Z, right?
Your assumption of adding R is part of the FG is only correct if that R is placed inside the Function Generator and performs part of the circuit or the transfer function in union with the capacitive reactance Xc and inductive reactance Xl of that circuit.
But placing a series resistor R outside the function generator is really a current limiting for the FG and causing voltage drop across it as the load varies (with or without varying frequency). It's really that energy is lost in that resistor and not transferred to the Transformer. It must be why no one wants to use 5K Zp OPT with another big dump 5K resistor to waste energy (power). It also limits the current flowing through the 5K series primary.
It gets much simpler when you define a voltage source and a current source and an AC impedance source.
What can AC impedance source supply to a short circuit? Voltage drop across it's own internal resistance or impedance value and current proportionately to that resistance or impedance value right?
What can AC impedance source with 5K series resistor as in your suggest supply to a short circuit? Find that out for me and let me know. Where does the voltage drop occur at with respect to Ground?
.
.
.Thou shall not stand where I type for I carry a bottle of Certified Audiophile Air and a Pure Silver Whip.
So, you are saying whatever resistor connected to the source becomes the actual source including the resistor value?
yes.
50ohms impedance with 1K series is 1050ohms source and 5K becomes 5050ohms source?
yup.
So, why don't we just all use 5K ohm OPT in series with a 5K ohm resistor and call it a true 10K ohm transformer?
It will provide a 10K load to the source but the problem is 1/2 the voltage will show up across the resistor and 1/2 across the transformer primary netting only 1/2 the voltage being delivered to the load.
Your assumption of adding R is part of the FG is only correct if that R is placed inside the Function Generator and performs part of the circuit or the transfer function in union with the capacitive reactance Xc and inductive reactance Xl of that circuit.
really? pit the resistor in the box and it suddenly disappears? My signal genrator has a variable output impedacne (50, 75, 135 and 600r) and it is accomplised by a switched resistor network at the output of the device. Sure those resistors are in the box but behave no differently that a similar resistor outside the box.
But placing a series resistor R outside the function generator is really a current limiting for the FG and causing voltage drop across it as the load varies (with or without varying frequency).
that is pretty much the definition of output impedance.
It's really that energy is lost in that resistor and not transferred to the Transformer.
again that is what output impedance does.
It must be why no one wants to use 5K Zp OPT with another big dump 5K resistor to waste energy (power). It also limits the current flowing through the 5K series primary.
If you drive a 5K source from a 5K load you lose 6dB of voltage.
It gets much simpler when you define a voltage source and a current source and an AC impedance source.
then you just need to understand how to implement what you define.
What can AC impedance source supply to a short circuit? Voltage drop across it's own internal resistance or impedance value and current proportionately to that resistance or impedance value right?
right so i you add series resistance to increase the source impedance ohms law dictates that you must also limit current. "current limiting" is not a function of added series resistance it is the result of it.
What can AC impedance source with 5K series resistor as in your suggest supply to a short circuit? Find that out for me and let me know. Where does the voltage drop occur at with respect to Ground?
it can source the exact same thing as if it had a 5K output impedance and no series resistor.
dave
Putting a resistor or resistor selection switch inside the box does not make it disappear. It makes it part of the circuit inside the FG and reacts along with other components inside that circuit.
Putting a series resistor outside the FG puts a series or parallel resistance to the load and I always look it as part of the load.I don't know what is used inside the FG for switching impedance values but let's say it used series or parallel resistor of different values.
Let's say it uses 550ohm resistor for 600ohm impedance, do you can think you can substitute a 550ohm resistor inside that FG with 9950ohm resistor to make your FG look like 10K source impedance?
I would love to have something like that.
.
.
.Thou shall not stand where I type for I carry a bottle of Certified Audiophile Air and a Pure Silver Whip.
Edits: 09/04/16
Putting a series resistor outside the FG puts a series or parallel resistance to the load and I always look it as part of the load.
it is part of the load to whatever is driving it. (whatever is upstream) and it becomes part of the source impedance for whatever is downstream.
Let's say it uses 550ohm resistor for 600ohm impedance, do you can think you can substitute a 550ohm resistor inside that FG with 9950ohm resistor to make your FG look like 10K source impedance?
yes... that is how it works.
I would love to have something like that.
this whole discussion has stemmed form your refusal to accept this simple fact so you can either accept that you already can do it or deny it and continue to redefine electrical theory.
dave
Firstly, there are a lot of Dave Slagle on the world so I wanted to make sure you are whom I think you are.
Are you from Intact Audio who makes Transformers?
Secondly, I don't redefine anything. I followed what I learned in school many decades ago and do either AC Analysis or DC Analysis completely (not both together as I see fit).
If there is anyone redefining it, it is you who insists upon it as the convenience of simplifying circuits to your theory and advantages instead of the analysis and theory that is taught and learned by millions of E.E or E.T around the world.
Thirdly (or) lastly, I now understand where your misunderstanding and confusion arises from.
I attached a couple of pics taken straight from the schematic of your WaveTek 288 user manual downloaded from the BAMA site.
Please tell me exactly how these attenuators are used in the circuit?
Are these series resistors part of the circuit or not part of the circuit (outside of the circuit even though they are put inside the box) in your WaveTek?
Regards,
James
.
.
.Thou shall not stand where I type for I carry a bottle of Certified Audiophile Air and a Pure Silver Whip.
you have the correct dave slagle....
Please tell me exactly how these attenuators are used in the circuit?
Are these series resistors part of the circuit or not part of the circuit (outside of the circuit even though they are put inside the box) in your WaveTek?
for attenuation closing K1 shorts out the π attenuator and for output impedance it appears to me that K2 and K3 either add or remove 25 ohms or 550 ohms in series with the output which coincides nicely with the selectable 50, 75 and 600 ohm output impedance. Now are you insisting that since these added series resistors are on the upstream side of J19 keeping them inside the box they behave any differently than if they were added on the downstream side? Or are the designers at wavetek now the ones taking liberties?
dave
The WaveTek Engineers are in the right and did exactly what I thought before too. They put these resistors inside the circuit with a closed feedback loop (R79) to the compensator circuit and microprocessor circuit which corrects any phase and voltage inaccuracies and send the compensated signal through the power amplifier circuit...Even then, they know the compensator circuit can correctly predict and correct the frequency below 1MHz so they put a lockout function for the higher impedance setting above 1MHz on the unbalanced output J79. It seems to me like a pretty good implementation to me.Read the user manual yet?
.
.
.Thou shall not stand where I type for I carry a bottle of Certified Audiophile Air and a Pure Silver Whip.
Edits: 09/05/16 09/05/16
From my understanding R79 simply allows for the sampling of the output voltage for the self calibration feature and is only used when the calibration cycle is running.
Indeed they do lock out frequencies above 1mhz when the larger resistors are used but we are talking about sub 100hz signals in this case and only looking for a rough estimation of actual in circuit behavior to decide if there is merit in actually lashing it up. If the transformer doesn't meet the desired low frequency results with a 5K series resistor there isn't a chance it will do better when it is loading a 10.
Applying rationale that may be extremely important in the megahertz+ range to this situation is probably not the best use of anyones time here.
dave
I presented you two pics together so you can connect the dots and at least come up with knowing the resistors are part of closed circuit with the feedback resistor of some sort going back into the more intelligent circuit compromised of the calibration circuit, microprocessor and the driving circuits.Instead, I got rebuttal and you come up with another explanation of what you think it is.
Anyway, observe the attached pic carefully including the V levels on the scope and explain to me what's happening at the transformer.
Have you seen anything like that before?
.
.
.Thou shall not stand where I type for I carry a bottle of Certified Audiophile Air and a Pure Silver Whip.
Edits: 09/05/16
I presented you two pics together so you can connect the dots and at least come up with knowing the resistors are part of closed circuit with the feedback resistor of some sort going back into the more intelligent circuit compromised of the calibration circuit, microprocessor and the driving circuits.
you want me to connect the dots you see and at no point do I see any mention or reference that that 100K resistor is a feedback resistor. I do see it connects to a "black box" called the "internal calibration network circuit" and know that the 288 has a calibration button but that is about as far as I get.
as for your measurements.... I have no idea what I am looking at. first the pictures are not very clear and the numbers are hard to read but you state that you used a 20V signal (didn't mention whether it was rms peal or p-p) and then show what appears to be a 40V p-p sine so that makes some sense. the other trace voltage label is not clear but it does appear to be 40mv which is a 1000:1 ratio or 60dB of attenuation and the only way I can even conceive of that happening is if the primary of the DUT is shorted.
Since you have gone as far as hooking up the circuit, why haven't you done the actual sweep with and without the 5K series resistor and shown the results?
dave
Don't worry about whether you see it or not. But now you know that the series resistor you see inside your FG is not quite the same as "Series Resistor" that you conveniently add to change impedance of the FG as you see fit.
If you still have in doubt, you can ask any Engineer about it.
If you add an external resistor to the source (FG, Amp, Preamp, Battery; etc.), it's just become part of the load circuit and does not do any impedance transformation to the source circuit. I guess it only took about 60 posts to get to this point.
Without further ado, I'll just go back and thinker what tubes to use with my IT. I think it'll do a lot of benefits to all of us following this thread.
I know you and Tre wanted to help and I appreciate it a lot. But I simply can't accept the wrong information presented for the sake of Engineering.
No hurt feelings meant to anyone!
.
.
.Thou shall not stand where I type for I carry a bottle of Certified Audiophile Air and a Pure Silver Whip.
"If you add an external resistor to the source (FG, Amp, Preamp, Battery; etc.), it's just become part of the load circuit and does not do any impedance transformation to the source circuit. I guess it only took about 60 posts to get to this point."
It seems t me you are misunderstanding a basic point about equivalent circuits. What Dave Slagle was saying seems to be correct. A source of 50 ohm impedance, in series with a 5K resistor, can be viewed as a source of 5050 ohms impedance.
It doesn't matter, from the point of view of analysing the total circuit, whether you view it as a source of 50 ohm impedance, driving into a load of 5K resistance in series with the transformer, or whether, on the other hand, you view it as a source of 5050 ohms impedance driving directly into the transformer. It is the same circuit either way you view it.
But it is perfectly valid, and useful, to view the 50 ohm output impedance generator plus the 5K series resistor as a 5050 ohm output impedance generator. The circuit as a whole doesn't care whether the 5K resistor sits inside the box labelled "signal generator," or if it sits outside the box.
Chris
Nope! You are just as confused as Dave and mixing up AC circuit with the DC ESR.
.
.
.Thou shall not stand where I type for I carry a bottle of Certified Audiophile Air and a Pure Silver Whip.
"Nope! You are just as confused as Dave and mixing up AC circuit with the DC ESR."No, actually not. It is really just an application of the Thevenin theorem.
Chris
Edits: 09/08/16
Chris,
I have said Tre and Dave before and I will tell you the same.
Even though it is not the same for the engineering point of view, the calculation would come out the same for the DC circuit with purely resistive load. I have no argument about it.
The problem with the (FG Z + external series resistor = new FG Z) when the load circuit has reactive component and it will change depends on the AC source. The results aren't the same everytime a signal from the FG increases or the frequency increases or decrease.
Regards,
James
.
.
.Thou shall not stand where I type for I carry a bottle of Certified Audiophile Air and a Pure Silver Whip.
In principle, the 50 ohm output impedance of the signal generator might have some reactive component (frequency dependent) to it. But when connected in series with the 5K resistor, the non-reactive 5K will be completely swamping the 50 ohms, and so any errors arising from treating the generator plus 5K resistor as simply being a "new generator" with a 5050 ohm output impedance (essentially non-reactive, to good approximation) will be very tiny.
Chris
Yes, it is true when the FG+5K resistor is connected to purely resistive LOAD.But we can't do that when the reactive load is connected on the other end of FG+5K series resistor.
I don't think I can explain you any better than I did to Dave.
You need to see a FG as an amplifier with a feedback loop so the output impedance remains near constant as possible under different loads. Any drop in voltage and current and phase coherence is being compensated by the FG internal circuitry. That is something an external series 5K resistor can't do. And that series resistor is not the origin of the actual DC/AC current or source.
That's not even remotely the same or close to a true 5K output impedance AC source.
.
.
.Thou shall not stand where I type for I carry a bottle of Certified Audiophile Air and a Pure Silver Whip.
Edits: 09/08/16
"Any drop in voltage and current and phase coherence is being compensated by the FG internal circuitry. That is something an external series 5K resistor can't do."
But the whole point is that a signal generator with 5K output impedance will NOT compensate for voltage drop in the way you are saying. If it did, then its output impedance would be lower. The output voltage of a signal generator with 5K output impedance will drop to one half of its unloaded output voltage if it drives into a 5K load. That is what it means to have a 5K output impedance.
If instead the signal generator compensated by means of internal feedback circuitry, so that connecting the 5K external load caused less than 50% drop in output voltage, it would mean that the output impedance of the signal generator was actually, by definition, less than 5K.
A signal generator with unloaded AC output voltage V and output impedance R can be modelled by an ideal zero-output-impedance voltage source with AC voltage V, in series with a resistor of resistance R.
Yes, there are certain approximation issues concerning some possible reactive component to the output impedance of the generator. But I would think that any modern half way decent signal generator will have an essentially resistive (non-reactive) output impedance, at least at low (like audio) frequencies.
And as I said before, if one is making a 5K output impedance (or 5050 to be more precise) signal generator by putting a 5K resistor in series with a 50 ohm output impedance signal generator, the issue of whether the 50 ohm impedance is reactive or not will pale into insignificance when put in series with 5K of purely resistive impedance.
Chris
Chris,
You are contradicting some partial truth with some false information. The 5K impedance of FG has some (even if limited) capability to supply the voltage without sagging when a load with 5K or more load is connected (regardless of it's resistive or reactive).
The 5K series resistor external to the FG has 0 capability of supplying any load current to the 5K LOAD unless the 50ohm FG supply some current to it first. So, based on the impedance of the load, it is just dropping voltage (stealing from the source) rather than working in union with the Source.
Very different, yeah?
.
.
.Thou shall not stand where I type for I carry a bottle of Certified Audiophile Air and a Pure Silver Whip.
I think it wiuld be helppful if you were to give *your* definition of what a signal generator with 5K output impedance means. I mean, be quantitative and say exactly what measurement would you perform that would produce 5K as the answer?I believe I have given a conventional definition of output impedance, namely it equals the resistance of the load you have to drive in order to cause the no-load ourput voltage to drop by 50%
So what is your quantitative measurement that would allow you to determine the output impedance?
Chris
Edits: 09/08/16
It is a voltage source with an active components as the current / voltage provider (namely tubes or transistors with some passive devices such as resistors and Xc and Xl in a closed loop circuit).
.
.
.Thou shall not stand where I type for I carry a bottle of Certified Audiophile Air and a Pure Silver Whip.
No, I'm not asking about a narrative discussion about what is "inside the box." I'm asking what measurement you would perform, entirely "outside the signal generator box," to determine the quantitative answer for what its output impedance is.
For example, what external measurement would you make in order to derermine whether the output impedance of 5K claimed by the maker was an accurate claim or not? I.e. How exactly are you defining output impedance? How would you measure it?
Until you give your operational and quantitative definition of output impedance, it is not really possible to have a sensible discussion.
Chris
Chris,I don't need this discussion. You can read to the beginning of this thread if you want to know why. I am not the guy started it.
I think you'll also find the answers to your questions at the beginning of this thread.
Thanks,
James
.
.
.Thou shall not stand where I type for I carry a bottle of Certified Audiophile Air and a Pure Silver Whip.
Edits: 09/08/16
The problem with the (FG Z + external series resistor = new FG Z) when the load circuit has reactive component and it will change depends on the AC source. The results aren't the same everytime a signal from the FG increases or the frequency increases or decrease.
this is case with any source with a greater than 0 output impedance. For this particular application it is the fact that the source Z is 5K that we are using to our advantage in order to find the effect of the reactive component (primary L). The variation you speak of is exactly what we are looking for so we can work it against the other knowns to find the unknown (primary L)
dave
Dave,
From your theory to work, you will need to the DCR, and either the inductance and the capacitance at that frequency. You also need to consider which is predominant factor (i.e. primarily resistive, or capacitive or inductive). You will also need to consider the phase angle based on the predominant factor. And then, you will need to figure out the amount of reflection to the source. And once you know that, that's the result for the one fixed setting (such as 2Vp-p, 20Hz). Once you change a little value, you are required to recalculate once over again.
Again, the source is the TRUE origin of the current / voltage.
I know what you have been saying all along on this topics. I wish you are right and things are that easy too. But it's not and I learned that very long ago. In fact, all of us (my classmates and I) got it wrong at the very beginning and puts a big smile on the Instructor face. Some good memory there.
Regards,
James
.
.
.Thou shall not stand where I type for I carry a bottle of Certified Audiophile Air and a Pure Silver Whip.
You seem to be completely missing the application here. The ultimate function generator in this particular situation is a tube with an Rp of 5K amplifying the random signals we call music. This is what needs to replicate for a precursory test.
You mention your ideal for a function generator being an amplifier with a feedback loop yet this actual application doesn't meet that criteria so I guess we now need to wrap feedback around our tubes to make them work properly.
Back to the original task, the 5K series resistor will work perfectly for this application. It may show issued in different applications however this is not about designing a universal device it is about testing a very specific application. The main thing of interest is to determine the primary inductance of the DUT to see what tubes it is suitable for loading across the audio band.
dave
Dave,
Please don't start on the topics of amplifiers and feedback? It took me a good part of a year on that subject learning all types. It definitely doesn't serve the purpose of this thread.
The tube with Rp of 5K-10K is what I want to use as a source to test the transformer for an actual bandwidth and FR curve.
I have very little interest in the actual L value or the Z value but a nice HP Impedance Analyzer would be an addition to my toys in the future.
Regards,
James
.
.
.Thou shall not stand where I type for I carry a bottle of Certified Audiophile Air and a Pure Silver Whip.
Adding a 5k resistor to 5k transformer dosnt make it a 10k transformer because the winding ratio is still the same. You just add extra loss to it.
Obviously, that is exactly what it is! Extra Loss in the dumb resistor.The voltage drop (and power wasted) on the series 5K resistor isn't always going the same as the 5K transformer though. It could be more than half or less than half of the Vsource depending on whatever actual AC impedance of the transformer is.
A series 5K resistor with a 5K Transformer load is a pretty dumb way to make an amplifier FR curve goes crazy which would otherwise can be very good looking FR curve with a real 10K transformer load connected to it.
.
.
.Thou shall not stand where I type for I carry a bottle of Certified Audiophile Air and a Pure Silver Whip.
Edits: 09/05/16 09/05/16 09/05/16
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: