|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
173.189.219.68
Finally got a break in all my spring and summer chores and got a chance to work on the 26/26/45 amp again.
Decided to breadboard one channel again and then to try and apply all that to my real amp when it was working properly. Everything went just fine except for one major flaw that really has me baffled. I first put in just one 26 and got that working and then added another 26 (see schem) and got that working too. Only one catch, it seems to sound better and have more gain with just the one 26. Sounding better might make sense if it needed tweaking, but more gain? That seems counter intuitive.
Anyone got any ideas what is wrong with this design that would cause that?
This is my version of Drummer Will's amp, but the main difference is I used a second coupling cap between the second 26 and the 45 and Willie's is direct coupled. That's not likely to cause what I am seeing is it?
Edits: 05/24/16 05/24/16 05/24/16Follow Ups:
I now have an all AC filament heated 28-26 DC 45 amp sounding great.
I found I did not need a cap between the two 26 tubes. and I DID NOT bypass the Rk on the first 26. This pretty much follows all the twin input tube sections I found in the archives.
Again here is a copy of my public Domain schematic for my working 26-26 DC 45 stereo amp.
Have fun
Willie
Willie,Choke off your rectifier is listed as 320 mHY or 0.32 Hy. If its a Traid, it must be a C-40X and NOT a C-56U as you have denoted on your schematic.
WHICH Traid is it ???
Triad C-56U as I recall is used for " flywheel " supplies ala Pasternak, which is 35 mHY at 2 A. and 0.79 Ohms.
Your notations on "D-Link" caps needs to be corrected to "DC-Link" caps, so people understand. eg :
http://www.vishay.com/capacitors/film/dc-link/
WIMA makes em too.
Where IS Henry Pasternak, Ray Moth, and the ole gang of well-wishers ??
Lovely to my eyes, ( and your ears most likely ), to see you using all sub 10 Ohm DCR Ls in your filtering / supply DrummerWill. You need to teach Vinnie that technique.
Do all 'de Forum members notice you used doubled-up resistors on ALL plates and cathodes, to get the most dynamics possible...closely matched to each other ???
This build technique was likely first seen decades ago in the KILLER MFA Luminescence RIAA ( 70 pound ) preamps. A high-water-mark I'm told in preamp design and performance.
Jeff
Edits: 05/26/16 05/26/16 05/26/16
I am confused Willie. You say you did not need a cap between the 26's, but your schem shows one. Or is this not the schem for that amp?
I finally do have mine working again, but now I need to do some tweaking.
Switching over to dc heating on my 26 preamp allowed me to use ac heating on the amp filaments with virtually no hum. Made a big difference.
Might have to try some of your changes too. Sounds interesting!
Edits: 05/26/16 05/26/16
Oops.... I knew I should have looked at my own schemo , before I tried to tell someone what to do. You are correct, only DC to the 45 !...
Dear Dr. LowMu, the L1 choke is a Triad C-56U as labeled. My mistake was designating it as a 320 mhy choke, it should be labeled as 350mhy. Thanks for pointing that out.
Have fun !
Willie
Triad C-56U is only 35 mHY, so its the start of a Henry Pasternak Flywheel ...fine !!
Rated DCR is only 0.79 Ohms, Current rating is 2 A.
Jeff
With an 880 VCT tranny the choke loaded input with SS rectification would be 0.9 time half the secondary winding or 440 times 0.9 or 396 VDC provided the choke meets critical inductance which is L = (B+/Iq)/1130
moving things around:
10*1130 = B+/Iq
Looking at the spec sheet for the 5R4:
ya might be OK. Maybe....
Well see before too long.
I finally found the problems, and there were two of them. While trying to trace a signal through the amp with my scope, I did not see any amplification in the signal coming out or the second 26. Upon looking around that circuit for mistakes in wiring, etc, I noticed that the filaments for that tube were not glowing. Checking the connections to the filament transformer I found that one of the clip leads had worked lose and was not making contact. No wonder I could not hear any difference between two 26's and only one. After I go that fixed I cranked it back up and noticed the B+2 for the second 26 was quite a bit lower than the first 26. Tracing through that circuit I found that I had taken the b+2 voltage off the wrong side of the resistor reducing the b+2 on the first 26 and so was getting a much reduced b+2 that I was further reducing with the resistor for the second 26. Got that all straightened out and now everythsing seems to be working as it should. The gain is much improved and it sounds pretty good. Now I will have to go back and plus everything up to the operating points we came up with the first time around, and then maybe I can try incorporating these changes into the real amp. Thanks for all the help folks, it got me looking in the right direction.
Edits: 05/25/16 05/25/16 05/25/16
IMO/IME, You would be better off going for a DC 45 final circuit and using something like a Triad C56-U as an input choke to get the B+ a bit higher.For the cap coupled final, you will need to drop more voltage off the rectifier somehow for lower B+ for the 45 plate, then decouple for the 26 driving stages.
I have done that before with a big power resistor and cap off the rectifier. Or with a power resistor in series with the primary HOT lead on the PT.
The major issue is the HV secondary of the PT in your circuit.
It is better suited for a DC 45 final.
Do you have another PT with a lower HV secondary voltage?
Edits: 05/25/16
I have a feeling by the time I get both channels of this thing operational the B+ won't be too high anymore.
Right .
A second channel should drop the B+ a good bit.
vinnie2,
OUPUT STAGE
You B+ for the 45 is WAY to high. The maximum plate to cathode voltage for the 45 is 275. Now add about 50 volts of cathode bias and 275+50 is 325 VDC for the power supply. You have 400! Further, you need a cathode bias resistor. For 35mA, this is 50/35mA and I generally use 1.5k Mills MRA-12. The bypass cap is approximately, C=1/(2*pi*R*F), so with 1.5k and f of 5 Hz, that's something like 20uF. I often use 47uF/400V Panasonic TSHA.
NOW, why did you get hear? First, the schematic you tried to copy was direct coupled. This is where the too much B+ came from. Second, you either forgot the cathode resistor which would have been much bigger or there were stacked supplies.
INPUT STAGES
You MUST remember that when you add an addition stage, that is one versus two 26 stages, you have to change the absolute polarity. You really can't make a comparison without doing this.
Each of the 26 stages inverts 180 degrees as does the output stage. The output you used is phased correctly for two stages but not three.
POWER SUPPLY
Generally, for a three stage amp you need not decouple the second stage from the output stage but you must decouple the input stage.
Do these things and then chase what others are saying....otherwise, you'll end up chasing your tail.
Hope that helps...
output
I'll have to check on the b+ fore the 45. It may be too high. I do have a cathode bias resistor, the schem is drawn wrong. As I said in a earlier post in this thread, the cap and resistor are actually below the pot not above it. Guess I had better correct the schem and repost it.
Input
This came up the first time I breadboarded this amp. I think Tre' brought it up. The preamp I am using has a single 26 tube so we decided that made things ok again.
Power supply
Are you saying I should take out the cap and resistor between the 26 and the 26 and 45 and make it direct coupled?
POWER SUPPLY
Nope. The B+ feeds the 45 and the second 26, then you decouple using a resistor and a cap to ground to feed B+2 to the first 26.
The way I have it right now I am taking the B+2 of the decoupling node for both 26's. Is that ok?
vinnie2, you gotta remember that the output stage and the second 26 stage are 180 degrees out of phase when drawing from the supply whereas the input or first 26 stage and the output are in phase. Add the difference in signal levels and this why you want to decouple the input stage. The tail will wag the dog, so to speak.
Decoupling both input and driver stages generally increases the supply impedance of the driver stage and doesn't sound as good as just decoding the input stage.
Is it necessary to decouple if the second stage is choke loaded or IT coupled from the plate of second stage?
Not sure I follow what your asking...
I'll have to take your word for it. After I got things working again (see update), I tried it both ways and was not able to hear any real difference. Of course I did not listen very long and I did not try a lot of different types of music, so it may be there. I will have to remember that if after I finish building it I want to make some mods. Besides that my ears are 69 years old now, and who knows what they are not telling me! I left it without decoupling just in case it does make a difference.
Is it wired as drawn?
Cheers,
01.
"Confusion of goals and perfection of means seems to characterise our age." Albert Einstein
I thought I had corrected the schematic! The resistor and the cap are actually below the pot, not above it as drawn. Sorry.
.
"Confusion of goals and perfection of means seems to characterise our age." Albert Einstein
Turn your pot up full and measure the AC voltage gain of both 26 stages, then compare that to the gain of the 26 stage closest to the finals.
If you got an AC signal generator and scope, that should be rather easy.
You should have more gain with both 26 stages active.
Thats true. If I put a sine wave through the amp it should be bigger after the second 26 than after the first. I will give that a try tomorrow.
400VDC seems kind of high for a non-DC coupled 45 final.
What is the Vp-k and bias current?
I run my 45's around 275Vp-k @ 37mA.
275 VDC times 0.037A. is 10.175 Watts continuous dissipation for a tube whose maximum plate dissipation is 10 Watts. No need to hotrod the Type 45 tube, dissipation-wise, and KILL tube life as you are doing it !!!
Consider running tubes at their Golden Ratio percentages of dissipation, for HUGE tube life, and beneficially, a LESS THERMALLY STRESSED sound to your ears.
In Golden Ratio terms, a 10 Watt MAXIMUM plate dissipation Type 45 may be nicely run at about 62% ( Golden Ratio percentage ) or 6.2 Watts.
A suggestion, run 45s at 250 and maybe 25 mA. which equates to 6.25 Watts. It will be less stressed sounding, and IF your wiring and supply is good, no loss of performance will be had.
An analogy, your car may be capable of running at 100 MPH, but it will do CONSIDERABLY better running it at 62 MPH continuously !!
Have fun, I am.
Jeff Medwin
Going to check on this. Got the same comment above.
When it sounds lousy, go through and measure/write down voltages. B+, plate, cathode, grit to cathode.etc... JH
Doesn't really sound lousy, just not as good as with one 26. Not a bad idea though; I will give that a try.
Edits: 05/24/16
Maybe there is some ultrasonic oscillation when two tubes are in (i.e. undesired global positive feedback through power supply).
From a theoretical point of view only two AC coupled stages are unconditionally stable.
ecc230
Paralell another 100uf with the 47uf on b+2 to the 1st 26. Thats a pretty high impedance supply feeding the tubes....even as low agm as they are. JH
Worth a try. We'll see what happens.
Forgot to mention the fact that when I had it breadboarded before I thought there was quite a bit more gain with the second 26. Either there really wasn't, or I have done something different (wrong) this time.
Doesn't seem to be any oscillation I can hear, It's just not as loud with two, and does not sound as good.
A low cost used EBAY 20mhz scope is all you need for this work. Not only can you see RF oscillation, you can also measure gains exactly. As for test tones, a PC sound card and free software is all you need for that or you can also buy a used audio signal generator.To spite what some people may say here, appropriate test gear is required to properly design audio equipment - as you are finding out.
Edits: 05/24/16
Actually I have both a scope and a signal generator, which I normally use for trouble shooting by tracing a signal until I find where it stops and I know where the problem is. This situation is a little differnt since the circuit is actually working, just not quite as good as i think it should.Are their any web sites that you know of that describe how to use them for the other things you are talking about? I haven not used them for that. I mostly build other peoples designs, so I use my test gear for trouble shooting.
Edits: 05/24/16 05/24/16 05/24/16
I don't know of a specific site for scope basics but I can imaging there are quite a few.
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: