It's the DynakitParts.com kit (current version) with all of their upgrades including the power supply choke and the individual bias board. Yes, they moved the negative feedback from the 16 ohm tap because the new transformers are 8/4 ohm and there is no 16 ohm tap The NFB component values have been checked and double checked by DynakitParts, and D. Gillespie agrees with the adjusted NFB part values that were chosen for the 8 ohm tap.
Weak bass could be a symptom of mismatched tubes DC saturating the transformer. Only adjusting bias is not
enough as different Gm ( amplification) will cause a DC component at power.
enough as different Gm ( amplification) will cause a DC component at power.
I’m a little surprised. Before listening to any amp, testing at least the voltages with a good DMM is important. Or else you could be chasing your tail.
But I am a testing freak. I measure all resistors, caps, passive parts; measure all transistors that I can, get tubes from reputable sources, and even test the output transformers for proper bandwidth, etc…and then put the enchilada together and test with my scope. Distortion tests are of use too and I’m on the toad to doing that.
Then comes the listening. In the process you end up understanding more of why/how the circuit performs.
But the water is warm…jump on in and enjoy the journey.
Best,
Anand.
That's a good thought. All I can say about that is that both the 6P14P and 6P14P-ER came as matched quads from reputable sources with the values written on the boxes. The 6P14P sounds great in the cheap Chinese kit that I built, so they can't be too mismatched.
I agree. It didn't smoke when I turned it on. It was late at night. I plugged in cheap speakers and it made sound. Into the system it went. No buzz or hum. Played music. That was that. 😉
Next step: now that the amp is out of the system, test all voltages per the voltage chart.
Last edited:
You mentioned something about your service manual being corrupted or unreadable. Many service manuals from old tube amp designs are available online and google is your friend.
See here: https://www.thetubestore.com/dynaco-schematics
Best,
Anand.
See here: https://www.thetubestore.com/dynaco-schematics
Best,
Anand.
Yes, it's a consequence of people running OCR I think. From the copy there at that site, the number 9 is missing from line 9 so I easily overlooked it. Also, the labels for the quad cap are missing.
From the current ST-35 manual from DynaKitParts.com it should look like this:
I have the amplifier upside down with the bottom off ready for testing, but right now the Labor Day BBQ requires my attention! 🥩🐔🍗🔪
Anything alarming in these results? This is the JJ 12DW7 drivers on V1 and V4 and 6P14P ouput tubes on V2+V3 and V5+V6. Why pin 1 of the ouput tubes varies so much is beyond me. It's an unconnected pin. I didn't measure the heaters.
Isn't this an error in the specs posted online? 215V vs. 230V left-right on the drivers? Which is correct?
If I use 215V for both I get this:
If I use 215V for both I get this:
Last edited:
All your voltages appear pretty close to target. Remember tube stuff can vary 5-10%.
With regards to the Dynaco printed spec, it could be a small error or what they actually measured due to mains voltage variations, tube variations, etc…nothing to lose sleep over.
Back to what matters.
It would be good if you get a scope or somebody with one who can do a Bode plot analysis and measure gain and phase end to end. Max power, visualize square waves, etc…As well as measure with a distortion analyzer or similar. All over your head, I know. But you have to start…somewhere.
And if then, you don’t like the amp then there are more modifications or you can accept the sound as is and fool around with tube rolling, coupling cap rolling or roll in whatever flooby dust audiophile stuff you want. The world is your oyster in diy.
Best,
Anand.
With regards to the Dynaco printed spec, it could be a small error or what they actually measured due to mains voltage variations, tube variations, etc…nothing to lose sleep over.
Back to what matters.
It would be good if you get a scope or somebody with one who can do a Bode plot analysis and measure gain and phase end to end. Max power, visualize square waves, etc…As well as measure with a distortion analyzer or similar. All over your head, I know. But you have to start…somewhere.
And if then, you don’t like the amp then there are more modifications or you can accept the sound as is and fool around with tube rolling, coupling cap rolling or roll in whatever flooby dust audiophile stuff you want. The world is your oyster in diy.
Best,
Anand.
Last edited:
I have asked in three places here on DIYaudio over the past few weeks if there is anyone in Northeast Florida - Central Florida - Southeast Georgia - maybe Western North Carolina - who would be willing to do some tests, but so far nobody. Purchasing the test equipment and learning to use it is beyond where I ever intend to take this hobby.
The build manual says values should be within 10% of spec, and they seem to be within the range.
I find the situation puzzling considering that I really like the $400 cheapo Chinese amplifier kit that I built that also uses the 6P14P/EL84 output tubes and a triode input (6N1 in that one). It's a different circuit though, with no ultralinear, no positive feedback, and just a single single negative feedback resistor.
The cathode voltages seems a little low ( 12,5 vs 13.5V ), if the cathode resistor 380 (95 x 4) ohm on each tube ?
I did note that pin 3 on all four output tubes is low across the board ... 12.5 vs 13.5V as you said. I don't know why.
All resistors are new Dale resistors from Mouser and DigiKey that were tested to be much less than 1% different from spec before installation.
The original circuit used one 95 ohm resistor (R13 on the original schematic) for all four output tubes, which is different as you noted. Pins #3 from all four output tubes now go through the bias adjustment circuit instead, which is adjusted to 35mA per tube using the potentiometer for each tube. The full schematic for the current build, including the added adjustable bias circuitry, is in post #1. The 95 ohm R13 is replaced by the individual bias adjustment in this amplifier. Perhaps this add-on bias adjustment circuit needs to be changed? Could this low voltage on pins #3 of all four output tubes possibly be the culprit here? It seems peculiar.
All resistors are new Dale resistors from Mouser and DigiKey that were tested to be much less than 1% different from spec before installation.
The original circuit used one 95 ohm resistor (R13 on the original schematic) for all four output tubes, which is different as you noted. Pins #3 from all four output tubes now go through the bias adjustment circuit instead, which is adjusted to 35mA per tube using the potentiometer for each tube. The full schematic for the current build, including the added adjustable bias circuitry, is in post #1. The 95 ohm R13 is replaced by the individual bias adjustment in this amplifier. Perhaps this add-on bias adjustment circuit needs to be changed? Could this low voltage on pins #3 of all four output tubes possibly be the culprit here? It seems peculiar.
Last edited:
Assuming that a measurement as described below actually equates to 35mA:
BIAS ADJUSTMENT
Bias adjustment will be set at each of the four multi-turn on board 10K pots. Each pot is strategically positioned on the board respective to the positions of the four output rubes.
With power off, set meter to measure resistance (Ohms). Place black negative probe on either of the two board mounting screw heads. Place red positive probe on terminal post V1. Adjust potentiometer R9 until you measure 400 ohms to ground.
Repeat above procedure at R10 & T2, R11 & T3, R.12 & T4.
Power on the amplifier. Set meter to DC Volts. Place black negative probe on board ground reference screw. Place red positive probe on terminal post V1. Adjust potentiometer R9 until you measure 0.350 VDC (350 mV).
Repeat above procedure at R10 & V2, R11 & V3, R12 & V4.
The current setting of each EL84 output tube should now be 35 mA. This can be adjusted to your preference via the on-board bias potentiometers, providing a useable range of 320 to 450 mV (32 to 45 mA).
Note from Triode Electronics: Cathode current is 35mA. Plate current is about 30mA and the screen is 5mA. Bias all four tubes this way. Check again after a few hours use and readjust if necessary. Check again after one week, then one month.
Never Get Old,
Is your amplifier un-modified and completely identical to the ST-35 schematic in Post # 157?
1. The first thing I note is that All 4 output tubes use a single common self bias resistor, R13.
That was great in the old days (the tubes that came with the amplifier were very well matched for that exact circuit).
If you do not have extremely well matched output tubes, you will have Un-matched current in each output transformer's primary windings.
And, so . . . there goes the Bass, global negative feedback can not fix that (The ST-35 has global negative feedback).
Eurotubes.com has extremely well matched EL84 Quads.
For tubes that are not quite as well matched, then this will help . . .
If you can find the room, then convert the ST-35 to have Individual self bias resistors, and individual bypass caps across them.
The resistance of individual self bias resistors is 4X the resistance of R13.
Electrolytic caps are smaller than they used to be; and the resistors can be 1/4 of the power rating of R13 (save some space when multiplying the number of parts by 4X).
Just food for thought.
2. The second thing is that any parts that are not the same values as in the schematic, and . . . all bets are off to having the ST-35 sound like the exact ones that so many people like and do not complain about.
3. What loudspeaker model do you have?
What is the rated efficiency?
What is the DCR across the input terminals? (use your DMM to measure)
Let's get your ST-35 performing like what you want to hear.
Is your amplifier un-modified and completely identical to the ST-35 schematic in Post # 157?
1. The first thing I note is that All 4 output tubes use a single common self bias resistor, R13.
That was great in the old days (the tubes that came with the amplifier were very well matched for that exact circuit).
If you do not have extremely well matched output tubes, you will have Un-matched current in each output transformer's primary windings.
And, so . . . there goes the Bass, global negative feedback can not fix that (The ST-35 has global negative feedback).
Eurotubes.com has extremely well matched EL84 Quads.
For tubes that are not quite as well matched, then this will help . . .
If you can find the room, then convert the ST-35 to have Individual self bias resistors, and individual bypass caps across them.
The resistance of individual self bias resistors is 4X the resistance of R13.
Electrolytic caps are smaller than they used to be; and the resistors can be 1/4 of the power rating of R13 (save some space when multiplying the number of parts by 4X).
Just food for thought.
2. The second thing is that any parts that are not the same values as in the schematic, and . . . all bets are off to having the ST-35 sound like the exact ones that so many people like and do not complain about.
3. What loudspeaker model do you have?
What is the rated efficiency?
What is the DCR across the input terminals? (use your DMM to measure)
Let's get your ST-35 performing like what you want to hear.
Aha,
You have individual cathode self bias now?
What are the capacitance of the 4 bypass caps from each cathode to ground?
Too little capacitance, there goes the Bass.
What else has changed?
No output transformer want to see 10% difference in push and pull plate current, and some of them will not even take 5% difference.
Just a little difference can cause early saturation . . . there goes the Bass.
Saturation can also cause intermodulation with the mid and high frequencies.
You have individual cathode self bias now?
What are the capacitance of the 4 bypass caps from each cathode to ground?
Too little capacitance, there goes the Bass.
What else has changed?
No output transformer want to see 10% difference in push and pull plate current, and some of them will not even take 5% difference.
Just a little difference can cause early saturation . . . there goes the Bass.
Saturation can also cause intermodulation with the mid and high frequencies.
Correct. See schematic in post #1, with all values followed exactly per schematic. It has the add-on individual tube bias kit sold by DynakitParts, as shown on the schematic.
Compared to the original? A number of things. Thus the "series ii" attached to the name. See post #133 for a full list.
"Sounds" about right. Haha. Well, not haha for me I guess. 🙁
Four 470uF Nichicon Muse audio-grade electrolytics from Mouser, measured to within 0.5-1%.
Last edited:
Just an aside.
What do you drive your amps with?
-preamp ?
-direct (with pot or variable output) ?
-processor ?
What do you drive your amps with?
-preamp ?
-direct (with pot or variable output) ?
-processor ?
Never Get Old,
Your schematic in Post # 170 is not the same as it is in your Post #1.
R9 and R8 are in the control grid, g1 circuit.
They are not bias resistors, they are grid return resistors.
Is your circuit modified to apply separate bias voltages to R8 and R9?
That is completely different than Post # 1.
The cathode bias resistor networks in Post # 1 does not allow you to measure the current in the individual cathodes; because when you change the setting of the pots, the resistance from the cathode to ground changes, but is only measureable with the power off.
You need individual 10 Ohm resistors from ground to the bottoms of each bias resistor network; The cathode current is V across 10 Ohms/10 Ohms
350mV across 10 Ohms is 35mA.
From the multiple documentation of your posts, I feel like a dog chasing its tail.
So far, I am not convinced that you are running 35mA cathode current in each individual output tube (that is the sum of plate current and screen current).
Your schematic in Post # 170 is not the same as it is in your Post #1.
R9 and R8 are in the control grid, g1 circuit.
They are not bias resistors, they are grid return resistors.
Is your circuit modified to apply separate bias voltages to R8 and R9?
That is completely different than Post # 1.
The cathode bias resistor networks in Post # 1 does not allow you to measure the current in the individual cathodes; because when you change the setting of the pots, the resistance from the cathode to ground changes, but is only measureable with the power off.
You need individual 10 Ohm resistors from ground to the bottoms of each bias resistor network; The cathode current is V across 10 Ohms/10 Ohms
350mV across 10 Ohms is 35mA.
From the multiple documentation of your posts, I feel like a dog chasing its tail.
So far, I am not convinced that you are running 35mA cathode current in each individual output tube (that is the sum of plate current and screen current).
Last edited:
Your static DC voltages look fine, and your adjustment method is correct. Just can't expect less than 10% variation between design center values from the Nixon era and modern everything-new. That stuff is all much closer than should be expected.
Without some test equipment it's tough to plot the most likely next step, but I'd go for the twitchy things that get left off of cost constrained designs, firstly grid (and G2) stops and an output Zobel. If you can hear it but can't measure it with a DVM, decent odds it's either a gross wiring error (seems unlikely given your attention to detail) or parasitic oscillations. Grid stops and Zobels were often removed to save a little money, and at the time and with the exact construction specified at the time, it worked. Different parts, different construction, maybe not so lucky.
Personally, I wouldn't bother with the 7247 grids, and they're on a PCB, but the output valves could use (about 1K Ohm or so) grid stops, mounted right at the socket, and about 100R or so G2 stops, also right at the socket. A reasonable output Zobel for 8 Ohms tap is 10R, 1W or 2W, in series with 0u1F film cap. Can't hurt and might fix your issue. Shooting in the dark, but these steps have no downside other than cost and effort.
All good fortune,
Chris
Without some test equipment it's tough to plot the most likely next step, but I'd go for the twitchy things that get left off of cost constrained designs, firstly grid (and G2) stops and an output Zobel. If you can hear it but can't measure it with a DVM, decent odds it's either a gross wiring error (seems unlikely given your attention to detail) or parasitic oscillations. Grid stops and Zobels were often removed to save a little money, and at the time and with the exact construction specified at the time, it worked. Different parts, different construction, maybe not so lucky.
Personally, I wouldn't bother with the 7247 grids, and they're on a PCB, but the output valves could use (about 1K Ohm or so) grid stops, mounted right at the socket, and about 100R or so G2 stops, also right at the socket. A reasonable output Zobel for 8 Ohms tap is 10R, 1W or 2W, in series with 0u1F film cap. Can't hurt and might fix your issue. Shooting in the dark, but these steps have no downside other than cost and effort.
All good fortune,
Chris