It's a hard question to answer which is probably why no one wanted to weigh in.
I'm not familiar with the particular headphone amp, but I'm assuming it operates at rather low voltages? Could be that this particular tube due to brand/era/factory specific design quirks has higher transconductance at low voltages than is typical for a 12AU7A.
If this is really a genuine Tesla tube then it has to be considerably older than a decade at this point. You should try to acquire a few more samples of this tube quickly and determine whether the behavior is consistent or specific to this one tube.
FWIW if it works and sounds good to you I'd just use it.
I'm not familiar with the particular headphone amp, but I'm assuming it operates at rather low voltages? Could be that this particular tube due to brand/era/factory specific design quirks has higher transconductance at low voltages than is typical for a 12AU7A.
If this is really a genuine Tesla tube then it has to be considerably older than a decade at this point. You should try to acquire a few more samples of this tube quickly and determine whether the behavior is consistent or specific to this one tube.
FWIW if it works and sounds good to you I'd just use it.
Last edited:
Appreciate it, and it may be simple haha, but as I stated I am a noob to tubes. Correct, this is a low voltage amp that runs off a 24V power supply. Upon some research it looks like I was confused by the masses calling JJ tubes, tesla tubes. Just read up that they use the same old tooling and so forth but different companies. This is a JJ gold pin, not a tesla. Will get another tomorrow and take this one in to be tested and see what they say. Again input is appreciated!
Any chance you could link or attach a schematic, please? Would make it much easier to figure out what you are working with. 😉
I'd be sure to hold onto that tube even if they offer to replace as it's possible, but not entirely likely that it is an odd one.
Thanks! Will do on that. I am going to pick up another today and check it out. I actually hope it is not a fluke as I really like it.
I picked up this schematic off of rock grotto...
Located at this thread
http://rockgrotto.proboards.com/index.cgi?board=m&action=display&thread=5297&page=1
It is just a cheap ebay headphone amp that a lot of headphone enthusiasts toy with as it actually can sound really good with some mods / tweaks.
Somewhat surprised to see there is no provision for cathode bias in the 12AU7A - I guess that is what is necessary to get it to work on a 24V supply. (I've never designed anything to work on voltages this low.) I suspect a lot of grid current is flowing, can you confirm by measuring voltage drop across R4 or R5.. I will resist the urge to try and "improve" this design 😛 but I can see how would be very sensitive to very small parametric differences in the tubes - this is entirely outside of the normal set of operating conditions for a 12AU7.
The schematic does not follow either of the standard drawing conventions so I am wondering if the bases of the CCS transistors that load the plates of the 12AU7 are in fact connected to the cathode of LED1. I don't see any circuitry for adjusting the CCS current so do you have a variant with an additional pot? (The two shown are just volume controls.) Based on your comments it sounds to me like the JJ tube is biased pretty cold and there is not much headroom as the plate voltage is 20V and there is only 4V across the CCS. It may be more emission limited at this low voltage than some of your other samples or it may have higher grid current resulting in more effective grid leak bias on the grids. Who knows? 😀
The schematic does not follow either of the standard drawing conventions so I am wondering if the bases of the CCS transistors that load the plates of the 12AU7 are in fact connected to the cathode of LED1. I don't see any circuitry for adjusting the CCS current so do you have a variant with an additional pot? (The two shown are just volume controls.) Based on your comments it sounds to me like the JJ tube is biased pretty cold and there is not much headroom as the plate voltage is 20V and there is only 4V across the CCS. It may be more emission limited at this low voltage than some of your other samples or it may have higher grid current resulting in more effective grid leak bias on the grids. Who knows? 😀
Last edited:
I wouldn't expect it to be as there will be some section to section variation unless they are extraordinarily well matched.
Grid current incidentally is around 1uA. Is that voltage positive or negative relative to ground? I will say that this tube is not running anywhere near the linear portions of its characteristic curves - probably sounds good because you have fairly sensitive headphones that don't require a lot of voltage. (I have a pair of 32 ohm phones that are comfortably loud at 10mVrms so...)
It strikes me from a technical perspective that this JJ tube actually doesn't perform well in this circuit, but if it sounds good then let your ears be the judge. Try to get your hands on a NOS Mazda/Sifte ECC82/12AU7A as they are IMLE amongst the best sounding, but again this is at normal plate voltages..
Potentially bad advice 😱 : It seems to me that a similar design based around the 6GM8/ECC86 would perform a lot better.. (Running at say 10mA or so with a 39 ohm resistor in the cathode of each section..) With some judicious etch cuts and a few component subs this could even be done to this board. I'd do it.. 😀
Last edited:
Kevin,
The triode stage design is indeed odd, low voltage stages are usually done like you would expect, just with different component values. For grid leak bias R4 and R5 are of course way too small. Rectified signal bias?! 😀
Oh, and the ECC82 can indeed run at these low voltages, it is actually not that bad a choice. No need for an ECC86 or 10mA of current. Magnifying your datasheets helps. 🙂
Garage,
You need to help us here, we cannot read thoughts. 😉
*) In what way have you modified your amp compared to the schematic?
*) There is only one potentiometer in your schematic, R1. Is that the one you are talking about whem you mention adjusting 'the pots'?
*) What supply voltage do you use, and where did you measure the voltages you mention in your first post? (Up to 20V)
- Frank.
The triode stage design is indeed odd, low voltage stages are usually done like you would expect, just with different component values. For grid leak bias R4 and R5 are of course way too small. Rectified signal bias?! 😀
Oh, and the ECC82 can indeed run at these low voltages, it is actually not that bad a choice. No need for an ECC86 or 10mA of current. Magnifying your datasheets helps. 🙂
Garage,
You need to help us here, we cannot read thoughts. 😉
*) In what way have you modified your amp compared to the schematic?
*) There is only one potentiometer in your schematic, R1. Is that the one you are talking about whem you mention adjusting 'the pots'?
*) What supply voltage do you use, and where did you measure the voltages you mention in your first post? (Up to 20V)
- Frank.
Thanks guys. I will have this tube tested today and see what they say, I am not complaining at all about it as it sounds fantastic even in this little amp. It just behaves so different than my other tubes and was just looking for some insight on to the possible reasons.
Also my headphones are... 150ohm and the second set is 300ohm.
Also my headphones are... 150ohm and the second set is 300ohm.
The changes I have made are as follows...
- Changed the stock irf630's with irl530's
- Changed out the fixed 3K resistors at r13 and r14 with 5K pots
- Changed the stock 6800uF 25V at c7 to a 4800uF 35V as it did not make much sense to run a 25v cap on 24V input
- Changed out the two 1000uF 85C caps to 105C 2200uF at c4 and c5 since one of the units is right next to a scalding hot lm317
- Cut the traces between pin 4 and 5 as the filaments are connected which was the cause of a ugly crosstalk problem. Pic below shows the issue, cut traces in red and jumper wire in blue taken from rock grotto
- Changed the stock irf630's with irl530's
- Changed out the fixed 3K resistors at r13 and r14 with 5K pots
- Changed the stock 6800uF 25V at c7 to a 4800uF 35V as it did not make much sense to run a 25v cap on 24V input
- Changed out the two 1000uF 85C caps to 105C 2200uF at c4 and c5 since one of the units is right next to a scalding hot lm317
- Cut the traces between pin 4 and 5 as the filaments are connected which was the cause of a ugly crosstalk problem. Pic below shows the issue, cut traces in red and jumper wire in blue taken from rock grotto
Wow, you've done all the right stuff. 😀 I would however add a small resistor in series with the CCS pots to avoid saturating the CCS.
Things to try would include increasing the value of R4 and R5 to make the grid current work for you, this could develop some real grid bias. Anything up to 1M might be worth trying, but I suspect the optimum resistance value would probably be a couple of hundred K ohms at most. Cathode bias would be another thing to try with small resistance values..
You could try a 6GM8 with a minor revision to the filament wiring, but I suspect based on your comments about cross talk a largish cap across the filament would be required as well. So maybe that isn't the brilliant idea I thought it was.. 😀
Things to try would include increasing the value of R4 and R5 to make the grid current work for you, this could develop some real grid bias. Anything up to 1M might be worth trying, but I suspect the optimum resistance value would probably be a couple of hundred K ohms at most. Cathode bias would be another thing to try with small resistance values..
You could try a 6GM8 with a minor revision to the filament wiring, but I suspect based on your comments about cross talk a largish cap across the filament would be required as well. So maybe that isn't the brilliant idea I thought it was.. 😀
Last edited:
Well the mods I cannot take credit for, the guys over at rock grotto came up with them. I just can sling solder quite well so followed their lead 🙂
They mentioned a 1K resistor across the pots but I did not know exactly what it was doing / they said it was optional so did not include it. You think it is a good idea then? I will add those in when I get a moment.
Only one major thing left, Need to isolate hum in the unit. It is present with any tube but goes away when I touch anything on the unit that is groun or if I touch any ground on the unit with a wire that is connected to earth. I assume it is being caused by the cheapy switching power supply they included.
They mentioned a 1K resistor across the pots but I did not know exactly what it was doing / they said it was optional so did not include it. You think it is a good idea then? I will add those in when I get a moment.
Only one major thing left, Need to isolate hum in the unit. It is present with any tube but goes away when I touch anything on the unit that is groun or if I touch any ground on the unit with a wire that is connected to earth. I assume it is being caused by the cheapy switching power supply they included.
Last edited:
My little budget gem with the jj gold pin glowing away.
An externally hosted image should be here but it was not working when we last tested it.
Ok picked up another JJ 12au7 gold pin yesterday, same as the one I felt that biased strange / but sounded awesome! And the results? Same as the first! Cannot get it much below 16V but the right channel I had to raise the voltage a little to get it to match. Sounds identical to the first unit so I am very happy. Great little tube!
- Status
- Not open for further replies.