|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
24.168.43.210
This project started as a SE KT88 amp. All holes are drilled, chassis painted...As an afterthought I tested the power xfmr and found that I got 1120VCT instead of 1000vct that someone had labeled. Choke loading +losses would have given me ~ 410vdc. Now I anticipate ~ 490vdc (choke loaded). The ? is, now what; the original design is now in the scrap pile. What would you do with 490vdc and Tango XE 20s.
Follow Ups:
The most obvious answer.
With a 200 plus Ohm DCR power transformer, its in the spirit of things. Add an Ef XFR below deck. After all, the chassis is painted so nicely, and you have those beautiful meters to look at.
Jeff Medwin
This tube doesn't mind high voltages. Looks like you can run almost 500v on the plate, -18v on the grid, and 100ma dissipation. cheers, Dak
Thank to all for your comments. I'll hook up a load and see where it all lands.
Jeff...I knew you were gona say that. What took you so long. It's a good idea and will likely take that path if the KT88s don't work out.
Two stage direct couple it !!
Jeff Medwin
You haven't said what operating point you had designed for, or what the transformer impedance is, or whether you were looking at a triode-wired or pentode mode amp, so I can't address any of the possible circuit changes. So I have only two points to make:
1: as others have said, you are likely to see way more than 14 volts losses once the supply is providing current.
2: A tube rectifier such as a 5R4 will drop plenty of volts.
The KT88 was to be triode/UL switchable with a 24v zener drop to the screens in UL.
I figured the fast warmup of a 5r4 would create a V spike before the KTs warmed up. I'm using a 370Vac motor run cap paralled with 370V film snubber cap (total around 80uf). I was counting on the slow turn on of a 5ar4 to prevent the spike (which could go to 700 V). Though the caps may take the spike, the driver tubes might curse at me and quit.
With that much dcr it will end up dam close to 400vdc...with 150ma draw. Mock it up with clip leads and find out man! JH
You can use the PSUD code to simulate the power supply; it will give you a good estimate of the voltage and allows various tube rectifiers.
If you need to operate at different voltages than your circuit was developed for, then some design changes will be needed to get the best performance. These things can be calculated; here is one example:
In triode mode, running higher voltage would call for reduced current if you want to retain the plate dissipation. Both of those changes call for a higher impedance output transformer. Ultralinear also calls for a higher impedance transformer than triode, if you want to take full advantage.
You can use a 555 timer circuit with a relay to delay the filament voltage to the rectifier tube to ensure the rest of the tubes are ready to conduct when the B+ comes up. I have used this approach and it works well for me.
"It is better to remain silent and thought a fool, then speak and remove all doubt." A. Lincoln
Did you get 1120V measured unloaded. If so then you will find that when loaded you will get more like the 1000V on the label.
Cheers,
Ian
Thank you for your help.
I got 1200vct unloaded. I got 565-0-565v using a variac to decrease the primary volts to where the 6.3v taps measured 6.3v. Likely I'll get more V drop once loaded but it will still be too high.
"I got 565-0-565v using a variac to decrease the primary volts to where the 6.3v taps measured 6.3v. Likely I'll get more V drop once loaded..."
Not if the 6.3V taps were also measured without a load. That would mean your variac wasn't turned up enough when you measured the HV. In any event, I would do as Paul suggested above, plus use a critical value choke. There's no reason not to continue with the project as planned. If you're interested, post the values for LC components you're planning to use (including DCR), and I can look at the output Z over frequency with SPICE. Better to do that now than after the design has been finalized.
--------------------------
Buy Chinese. Bury freedom.
Thanx for commenting. The secondaries measure 141ohm to ct and 152ohms to ct or 293 ohms leg to leg. The choke is a Stancor swinging choke 76ohms dcr and rated for 400ma and 4Hys. The cap (after the choke) will likely be ~80uf or so (this can change easily). primarys for the Tangos are 100 ohms.
The original operating points for the KT88 were to be 400v at 75ma with fixed bias. This is no longer the case. New operating points and likely self bias are needed.
Likely I'll set up some kind of load for the amp within a week and measure my B+ at full load within a week.
" The choke is a Stancor swinging choke 76ohms dcr and rated for 400ma and 4Hys. ... The original operating points for the KT88 were to be 400v at 75ma..."
Sorry, I meant to get back to this sooner. Can I ask why you're using a swinging choke? The amplifier is Class A, so average current draw doesn't change much when the amplifier is driven. That means the choke won't "swing." In any event, it's necessary to know the approximate inductance at the actual operating current (150 mA) in order to analyze this. For the purpose of simply taking a quick look, I'll assume it's 5H (the 4H rating is at 400 mA). If this choke is followed by 80uF, the following curve represents relative levels looking back into the supply:
As you can see, the curve is nice and smooth, with no irregularities that might indicate resonances "visible" to the amplifier. Just for contrast, here is the same supply with a 10uF output cap:
In this case, insufficient capacitance creates a much higher overall output impedance. Attenuation at 1kHz is degraded 15dB relative to the previous example. In addition, resonance between the cap and inductor produces an impedance bump. The supply practically "lets go" at 25Hz, meaning it won't be able to properly "anchor" the amplifier. Amplifier output in this area will almost certainly be distorted and poorly defined.
--------------------------
Buy Chinese. Bury freedom.
Mr. Kingdom
Thanx for your looking into this. the graphs are helpful; I think I'll up the cap value even more to lower the imp in the bass region.
I did finally get it up and running with just the KTs running in self bias. First try got me 492v. Dialing the variac back so that 6.3 volts were on the filaments got me 472v, changing the rectifier to 5v4 got me down to 460v. with 45v self bias I'm now down to 415v across the KT88s. I can live with that.
To answer your question about using a swinging choke. It's true that the average i doesn't change much in class A. To my way of thinking, music in a class A circuit is a deviation from the mean; the louder the music the greater the deviation. Thus a swinging choke is needed. Also, SE circuits sound better to me with choke loaded supplies.
The most important reason for using a swinging choke is; it was on the shelf, taking up room, and needed to be used. A major factor in my designs is how to use up all the stuff I acquired in the Hamfests of the 1990's.
"I'm now down to 415v across the KT88s. I can live with that."
Excellent!
"To my way of thinking, music in a class A circuit is a deviation from the mean; the louder the music the greater the deviation. Thus a swinging choke is needed."
The deviation is instantaneous and doesn't change the average. The choke will only "swing" in response to average changes in current. This conversation does make me wonder though - will a swinging choke respond to instantaneous changes in current draw when those changes occur at a low-frequency rate? This needs more thought...
"The most important reason for using a swinging choke is; it was on the shelf, taking up room, and needed to be used."
I'm sure lots of commercial manufacturers have used that same design philosophy. :)
--------------------------
Buy Chinese. Bury freedom.
I see about 47v bias (triode mode), plus at least 40v lost in resistances. Not counting rectifier voltage drop. Should turn your 490v into 403 volts - sounds pretty close to me.
Edits: 11/23/15
I don't see a problem either. Once the amp is running that voltage is going to drop.
As long as you don't exceed the plate dissipation, the voltage is fine.
Dan Santoni
Paul, I do not see a problem here, depending on which specs you read, the max G2 spec is 500 to 600 volts, {this is the spec for the original tubes}
The current Russian and Chinese tubes have checkered information on their stoic ability. Some live up to, or exceed the original specs, and some blow up.
Ultra-linear is another possibility considering a KT88 tube was designed for THAT capacity.
The Mind has No Firewall~ U.S. Army War College.
Thank you for your post. SE KT88s should be run about 400v (within my knowledge realm) PP can go up to 550v or so. I may get more voltage drop than I calculated, but even 460V is a too high.
After I hook up a test circuit I'll know for sure what the DCV is and adjust accordingly or try something different. In the meantime I figured I'd shake the tree (ask for help) and consider what falls out.
Hello,
Are those interstage transformers in the back (2 square white items)? If so, you might want to look at Shishido SET options. Yields a straight through signal path (no caps) leveraging the interstage transformers. Lots of schematics here -
http://www.single-ended.com/Lagarto/shishido/shishido-skema.htm
No expert on these things, but if your transformer is yielding 490 v straight out, then the voltage doubler stages shown in the schematics might put you in the range of driving a transmitter triode. That would yield you a well powered SET amp - well worth having.
Regards -- Roger
Tango XE20s are too good for KT88s. This calls out for DHT finals. You might manage 300bs with 490v. That would be about 430v A-K, not impossible. But you might need to swap the mains transformer for one of the Hammond models with filament supplies suitable for a 300b.
Trouble is you don't have much room on the chassis if this is a stereo amp.
Thank you: Though the 300b is the obvious choice, and will be considered if nothing else works out, I was looking for something different.
what kind of power supply do you have? If it is CLC making it choke loaded will drop the B+ by almost 100v depending on your choke. regards, Dak
What if you make a LCLC power supply. You should be able to knock of another bunch of volts. Also if you use a 5r4 or 5u4 rectifier you would lose more voltage. cheers, Dak
The 5r4gy drops about 65-70V.
It's one of my favorite rectifiers.
Also, it's an ST glass, so if you go for looks that might be a bonus for you.
Janos
Thank you, but it has to be a IDH rectifier. Such as 5AR4, 5v4... a DHR will give a voltage spike of 750v for 10-15 seconds in this circuit.
"What if you make a LCLC power supply. You should be able to knock of another bunch of volts."
No, the first choke creates the RMS reduction (assuming critical value). The second choke will only lose voltage according to its DCR. Might as well use a resistor if that's the intent.
--------------------------
Buy Chinese. Bury freedom.
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: