Hoffman Amplifiers Tube Amplifier Forum
Amp Stuff => Tube Amp Building - Tweaks - Repairs => Topic started by: Leevi on February 24, 2013, 02:40:13 pm
-
I have built a Matchless clubman 35 according to the linked schematic.
http://www.el34world.com/Forum/index.php?action=dlattach;topic=597.0;attach=7002;image (http://www.el34world.com/Forum/index.php?action=dlattach;topic=597.0;attach=7002;image)
I wonder the very hot biasing of this amp. The cathode current is about 72mA per tube (in my case)!
The previous version with 6SH7 is running even hotter ~95mA.
http://www.prowessamplifiers.com/schematics/post/matchless_clubman.pdf (http://www.prowessamplifiers.com/schematics/post/matchless_clubman.pdf)
Should increase the cathode resistor from 330R in order the get the cathode current decreased
or is it OK to keep on that level?
BTW there is a fault in the attached schematic:
The filter cap with 220k should go to the first amplification stage ("E"),
6.8k to EF86 ("D") and 8.2 to PI ("C").
/Leevi
-
The cathode current is about 72mA per tube (in my case)!
And the plate voltage is?
-
I don't know if we can trust the voltages indicated on the schematic you linked. They imply that the screens of the EL34's draw 25mA per tube by themselves at idle.
But I recall my old Clubman running the output tubes very hot. It is perfectly reasonable for you to increase the cathode resistor to tame the idle current if you want. The original Clubman had 15w or 25w aluminum-finned wirewound resistors for the cathode resistors.
-
And the plate voltage is?
In my case 24v.
/Leevi
-
Plate at 24v :w2:
K
-
according to the 93 model schematic, the 25V/270R = 92.6mA per tube - 430Va-25Vk = 405Va-k x 92.6mAIk = 37.5Wa per tube.
is that be right? EL34 is rated at 25Wa max. and that's probably conservative, but 50% over spec?
perhaps there is a misinterpretation; what if the value of the 270R under each EL34 is really 470R? then Ib/EL34 is then 53.2mA and Pdiss/EL34 is then 21.5W.
in the 96 model, with the EF86 driver, Rk is 330R with an anode to cathode V of ~400V or ~33W Pdiss @ idle.
if the data in either scema isn't bogus then the clubman 35 is likely a very unreliable amp and/or the EL34s have a short life.
respectfully,
--pete
-
Plate at 24v
Sorry, plate voltage is 400v with 5U4GB and 430v with GZ34.
According to the data sheet of EL34 the maximum cathode current is 150mA.
So we are still below that.
/Leevi
-
What is the primary impedance of the OT you are using?
In this datasheet are two conditions for cathode biased PP:
http://frank.pocnet.net/sheets/129/e/EL34.pdf (http://frank.pocnet.net/sheets/129/e/EL34.pdf)
Because your voltage is so high you could try to double the impedance, that is if your OT is 3,4k.
-
So, with 400V plate voltage, 72mA cathode current, and 330Ω cathode resistor, idle dissipation calculates at 27 watts (max is 25W). That's ignoring screen current. If screens are drawing 6mA then idle plate dissipation would be 25 watts. You can check the screen current by measuring the voltage ACROSS the 1K screen resistor. 1v = 1mA, 5v = 5mA, 10v = 10mA, etc. Subtract screen current from cathode current to get the plate current.
You're probably fine as is. I'd probably change cathode resistors to 470Ω or 500Ω just to run a little cooler, unless it changes the sound drastically. I doubt it will though.
-
What is the primary impedance of the OT you are using?
It's 3.4k, I think OT with 4-5k primary would be better
You can check the screen current by measuring the voltage ACROSS the 1K screen resistor. 1v = 1mA, 5v = 5mA, 10v = 10mA,
It's ~20mA
/Leevi
-
is that be right?
I have stolen this question, and it is my new,,, sarcastic,,,ebonics question of the week.......thanks for that DL :icon_biggrin:
No offense meant to anyone who might take any,,,,just a comedic observation
Being a manager of grown men who act like children always keeps me searching for "new material"
-
I correct the value of the screen current.
It's 7mA/tube.
/Leevi
-
according to the 93 model schematic, the 25V/270R = 92.6mA per tube - 430Va-25Vk = 405Va-k x 92.6mAIk = 37.5Wa per tube.
is that be right? ...
perhaps there is a misinterpretation; what if the value of the 270R under each EL34 is really 470R? ...
That's why I'm saying I don't know if the voltages on the schematic are correct.
I promise you the original Clubman 35 did indeed have a separate 270Ω resistor for each output tube. And they ran the tubes hot enough to smell them cooking, though I didn't see any redplating on the one I had.
Somewhere, I have a hand-scrawled schematic for this amp from when I owned one, but I don't think I had the foresight then to write down any of the voltages. So I can't help there.
according to the 93 model schematic, the 25V/270R = 92.6mA per tube - 430Va-25Vk = 405Va-k x 92.6mAIk = 37.5Wa per tube.
But part of that cathode current is screen current. The schematic shows 405v on 1 side of the 1kΩ screen resistor and 380v on the other. 25v drop and 25mA in the screen at idle?
92.6mA - 25mA = 67.6mA
67.6mA * 405v = 27.4w plate dissipation, assuming the screen current really is 25mA
But 25mA * 355v (screen to cathode) = 8.9w idle screen dissipation
At which point the EL34 screens would have melted. So there's something wrong with the voltage numbers on the schematic.
-
So there's something wrong with the voltage numbers on the schematic.
I think they are close to right values.
Check for instance the 1996 schematic I attached. There is a table of measured voltages and according
to that the cathode current seems to be on level 82mA.
You're probably fine as is. I'd probably change cathode resistors to 470Ω or 500Ω just to run a little cooler, unless it changes the sound drastically. I doubt it will though.
I modified the cathode by changing the 330R to 560R. I got the following values:
Plate voltage 413V
Cathode voltage 27.7V
Screen current 5.3mA
Screen voltage 397V.
Plate dissipation 19W/tube
/Leevi
-
Vk = 27.7 Rk = 560 so Ik = Vk/Rk = 27.7/560 = 49.46mA
since Ik = Ia+Ig2 then Pa = Va-k/Ik = 417V - 27.7*49.46mA = 19.3W as leevi states.
But part of that cathode current is screen current.
yes, i do understand that. Ik is Ia+Ig2 but since Pa measured includes Pg2 then Pa = Va-k*Ik-Ig2 + Vg2*Ig2 i assume then the tube rating is 25W Pa + 8W Pg2 or 33W total. does that seem right? conversely the 6L6GC is 30W Pa and 5W Pg2 or 35W total.
Pdiss specs taken from mullard datasheets.
one of the conditions published in mullard datasheet is AB1 cathode bias with 450V B+ with 465R for each Rk with a 6.5KRa-a load Ig2 is stated as being 10mA per tube at idle and Ia is stated as being 60mA per tube at idle. Pout is stated as being 40W.
if then; Ik=70mA per tube with 450V B+ for Va and Vg2. then 70mA/465R = 32.55Vk. Pa = Va-VRk/Ik = 450V-32.5V/70mA = in other words, Pdiss(total) = 417V/70mA = 29.22W / tube. matcheless ran lower Ra-a with a slightly lower B+ but still idles within the 33W total / tube rating.
5.3mA for Ig2 for each tube seems believable based on what EL34 datasheet says for the similar condition described above - if anything leevi's amp seems to have Ig2 running conservatively.
leevi, if you measure the voltage drop across each screen resistor with your meter set to the lowest possible DCV resolution, the results are more accurate Ig2 numbers.
you could try 470R,390, or even 330R under each tube - bet it still sounds good as long as you don't let Pg2 run away.
respectfully,
--pete
-
I have a 1996 Matchless Clubman 35. Yes in deed they are biased WAY HOT. With the stock 270 ohm cathode resistors and a 5AR4 rectifyer it cooks the tubes( EL 34`s) at 145% max plate dissipation. Good tubes like NOS or the original Svetlana`s will last awhile depending how hard and long you push the amp. I put a pair of Sovtek EL34`s in and the plates glowed bright red within 2 seconds. I perfered to use GE 6L6GC`s with a 5U4GB rectfyer. I have since cooled the amp down with 470 ohm cathode resistors. Its not my main giging amp due to the weight with its massive transformers.
-
I found couple of 383 Ohm power resistors and installed them to cathodes.
Measured values are now:
Plate voltage 400V
Cathode voltage/tube 24V
Screen current/tube 7mA
Screen voltage 380V.
Plate dissipation 24W/tube
Cathode current: 63mA/tube
I think I'll keep that setup. The maximum Ia of EL34 is 150mA.
/Leevi
-
Plate voltage 400V
Cathode voltage/tube 24V
Screen current/tube 7mA
Screen voltage 380V.
Plate dissipation 24W/tube
Cathode current: 63mA/tube
How do you get 24W with those numbers? It looks like you subtracted the cathode voltage from the measured plate voltage to get 376v. That's the voltage between the plate and cathode and that's the number you need to use for calculating plate dissipation. You also need the plate current and that's obtained by subtracting screen current from the cathode current. That leaves only 56mA flowing thru the plate. So now, multiply .056mA by 376V to get idle plate dissipation equal to 21 watts.
I think you neglected the screen current. The whole reason for measuring the screen current was so you could subtract it from the cathode current to determine the true plate current. I generally ignore screen current for 6V6s and 6L6s since it's so small. But EL34s screens are hungry for current and should not be overlooked, especially when you suspect the tubes are idling too hot for your comfort.
But, it's just numbers. And Matchless was not too concerned with numbers. :grin:
-
The maximum Ia of EL34 is 150mA.
Keep in mind, that is the maximum cathode current at any instant; it particularly applies to the moment where input signal drives the grid to 0v and the plate voltage is pulled down by the load impedance.
For example, 25w/.15A = 167v which should show you it's not really a rating to consider directly to determine idle conditions.
-
How do you get 24W with those numbers? It looks like you subtracted the cathode voltage from the measured plate voltage to get 376v. That's the voltage between the plate and cathode and that's the number you need to use for calculating plate dissipation. You also need the plate current and that's obtained by subtracting screen current from the cathode current. That leaves only 56mA flowing thru the plate. So now, multiply .056mA by 376V to get idle plate dissipation equal to 21 watts.
Alright, I used the Weber bias calculator and didn't properly read the instructions there.
I would say 21W is better than 24W.
Thanks sluckey for the correction.
/Leevi
-
I am building an amp right now with a similar power section as the clubman. But I connected the cathodes together on the el34s. I was wondering if anyone would think this would be better. I was thinking the voltage would be more stable so I could use one 50u cap instead of two 250uF caps.
-
... I connected the cathodes together on the el34s. I was wondering if anyone would think this would be better. I was thinking the voltage would be more stable so I could use one 50u cap instead of two 250uF caps.
It's unclear.
If the amp stays class A (neither output tube ever shuts off), then a shared cathode resistor is fine (and maybe even leads to more-stable bias, as a drop in 20mA of cathode current in one tube is offset by the increase of an extra 20mA in the other tube). If you're pushing into class AB (one tube shuts off some part of the time), then bias may shift when one tube shuts off and the other sees a lower load from the output transformer. This is part of the reason for having a bypass cap across the cathode resistors.
Separate cathode resistors may also force Class A, at least as a limit: Imagine the tube receiving an input signal so negative as to push it near cut-off. If the tube could reach actual cut off (0mA of cathode current), then due to Ohm's Law there is also 0v across the cathode resistor (0mA * Any-Ω = 0v). 0v is the bias which forces the tube to draw maximum current, short of positive grid voltage & grid current.
So a shared cathode resistor could allow an output stage to stray into class AB with big input signals, because the lost current of one shut-off-tube is replaced/exceeded by the other tube passing more current. Since this can't happen with individual cathode resistors, perhaps they were kept separate to ensure Class A operation.
It then follows that the big cathode bypass caps are there to keep bias voltage stable by bypassing the variational current around the individual cathode resistor.
-
> If you're pushing into class AB ... then bias may shift when one tube shuts off
Gain shifts also.
Whether it "goes AB" or not also depends heavily on load. Since speakers are varying loads, and musicians are not impedance-technicians, loading is unsure.