I tried 3.3k, then 47k and 470k, the higher the resistance, the less wobble there is(But the wobble's amplitude rises)
When I used the chassis as star ground, the wobble would only appear when the potentiometer was set at a particular position, not 30 seconds after the tube starts producing sound.(Note that I moved the gridstopper resistors closer to the tube at the same time)
Hey, now that I think about it, I think I bought metal film resistors for the whole amp last year. Aren't tube grids not supposed to like them?
Is the tube really putting out DC? The amplifier acts exactly the same if I unplug the tube but not the 10kOhm.
unplugging the tube beakes the circuit to the 10k resistor unless you made a wireing error... what changes do you note in the sound with increasing the resistor...?
Don't go too high... the 33k is about the upper limit you should aim for
Notice how the inverted input of the amp is connected to the negative power supply line though the 10k resistor... the resistor and tube forms a voltage divider almost like the tube is another resitor... so depending on the operating point this may always be at a set negative voltage, or 0V... but it needs carefull matching with the tube stage... this is DC Voltage...(which has no sound- unless you count melting) and why you need the 3u3 cap between the 2 stages..
You realy should not be haveing such problems, the buffer is basicaly a tube and 2 resistors... unless your power supply is very noisy, we should not be haveing issues...
Don't go too high... the 33k is about the upper limit you should aim for
Notice how the inverted input of the amp is connected to the negative power supply line though the 10k resistor... the resistor and tube forms a voltage divider almost like the tube is another resitor... so depending on the operating point this may always be at a set negative voltage, or 0V... but it needs carefull matching with the tube stage... this is DC Voltage...(which has no sound- unless you count melting) and why you need the 3u3 cap between the 2 stages..
You realy should not be haveing such problems, the buffer is basicaly a tube and 2 resistors... unless your power supply is very noisy, we should not be haveing issues...
Well, I tried 47kOhm(The closest to 33k I have) the wobble is a little slower, and there seems to be a little more bass but that's it.
And a little note, with the 10kOhm pin3 is 0.88v higher than ground, and with 57k(10k+47k) it's 1.67v.
I think so, after reading the JLTi topic it seems that it's very important.
I have much less filtering than the design asks, but it appears that even the real design's pi filter isn't blocking all the hum with others who had the same problem.
As well as the hum, I pickup other appliances being turned on and off.
Those we can deal with and fix, and it will further improve once you put your amp in steel box...
Do you have access to an old/dead/obsolete PC PSU we can hack for parts? it has two high voltage caps in we need.
Do you have access to an old/dead/obsolete PC PSU we can hack for parts? it has two high voltage caps in we need.
Oh, thanks, I forgot this one.
I have a pretty old huge(You could fit three 5½" drives in the case) inefficient SMPS I bought for dirt at a surplus store(Buying the fan it has in it alone was more expensive).
I just opened it, and there are two 1000µF 200V(sandwiched between two PCBs...) That's pretty much all since most of the caps are 6.3Vs used for the 24Amp 5V rail. The thing has WIMAs and Panasonics in it! In fact, there's a whole (shielded) PCB just for AC filtering/rectification with tons of chokes, small HV caps and large resistors(Let's say they cared more about clean power than generic PSU manufacturers...)
I just found this in a topic about the SLCF version of the buffer(found in the real JLTi) :
Well well well... When I wire my speakers in series, the wobble disappears, so maybe I should put the .22ohm finally... I found a .15 ohm in the SMPS PSU.
OK, I just wanted to confirm you can get two 200v caps...incase we can't get it to run with the default power, we can build a voltage doubler too...
Notice all the other nice stuff in the old PSUs, TO220 pads, screws, and screw insulators... little heatsinks... I usualy saw those up to make TO220 heatsinks from... Also thouse wound inductors can be very usefull, don't throw them away...
Notice all the other nice stuff in the old PSUs, TO220 pads, screws, and screw insulators... little heatsinks... I usualy saw those up to make TO220 heatsinks from... Also thouse wound inductors can be very usefull, don't throw them away...
Just wondering... Is there anything to try now or I should place my Mouser order and start playing in the amp again only in two weeks?
I was planning to use it in two weeks away from home(There's a portability reason for the size), do you think I could run it without tube and 10kOhm resistor safely?
I was planning to use it in two weeks away from home(There's a portability reason for the size), do you think I could run it without tube and 10kOhm resistor safely?
Oh, I put a few 0.0047µF 250V caps I found in the PSU in parallel with every diodes and I think that I get less hum than before(I didn't power the amp up before this today, all I know it that the hum is not that bad).
As for the wobble, I found that it only happens once there is a certain amount of current required(With speaker resistors). It goes back to normal once there's almost no sound. I feel I could probably solve the hum with tons of filters and a separate rectifier, but not the wobble.
For that I just thought about 3 solutions:
1. An LM337HV/317HV-based ±35V regulator for the tube's power lines.
Problems :
- Additional heat
- I'd need heatsink insulation : My PSU has no mica insulation, live heatsinks everywhere, there's enough space in it to make this safe. The LM3x7HVs aren't available in total plastic versions.
2. Two other transformers, well filtered AC and a tube rectifier.
Problems :
- Lack of space
- Need lots of parts.
3. Same as #2, but instead of new transformers, a smaller DC cap at the input. No worries about wobble as it would be filtered out.
Problems :
- No space for an other tube socket in a logical-looking place.
- The parts needed. (Not a problem for the tube and sockets, but more for the caps, etc.)
- Sound is probably going to be impaired since the 10Hz flat filter is a design choice.
#1 sound the easiest (to me) as the regulators would be lying vertically flat on the main heatsink : No chances to block the pot or the tube sockets. The heatsink only rises 8°C over room temp with almost constant DC wobble at the output right now(One LM3886 at the time though).
As for the wobble, I found that it only happens once there is a certain amount of current required(With speaker resistors). It goes back to normal once there's almost no sound. I feel I could probably solve the hum with tons of filters and a separate rectifier, but not the wobble.
For that I just thought about 3 solutions:
1. An LM337HV/317HV-based ±35V regulator for the tube's power lines.
Problems :
- Additional heat
- I'd need heatsink insulation : My PSU has no mica insulation, live heatsinks everywhere, there's enough space in it to make this safe. The LM3x7HVs aren't available in total plastic versions.
2. Two other transformers, well filtered AC and a tube rectifier.
Problems :
- Lack of space
- Need lots of parts.
3. Same as #2, but instead of new transformers, a smaller DC cap at the input. No worries about wobble as it would be filtered out.
Problems :
- No space for an other tube socket in a logical-looking place.
- The parts needed. (Not a problem for the tube and sockets, but more for the caps, etc.)
- Sound is probably going to be impaired since the 10Hz flat filter is a design choice.
#1 sound the easiest (to me) as the regulators would be lying vertically flat on the main heatsink : No chances to block the pot or the tube sockets. The heatsink only rises 8°C over room temp with almost constant DC wobble at the output right now(One LM3886 at the time though).
Screw the LM3x7HV, the LM337HV is totally impossible to find... (But can it be solved with dual regulators?)
Hey... I connected both channels and I get uber-wobble and hum now. (The 5" cones move about .5cm on each sides...)
I disconnect one speaker, and it's gone.
I disconnect one speaker, and it's gone.
My appologies for not returning yet.. life happens as you know... will try to come back to you later... got to go to dental specialist now about wisdom tooth... 🙁.. I have not read your further posts.. just don't break thing so long... ok...
scanned through a few posts.. dont forget there are insulators you can use in the PC PSU on the heatsinks... just unscrew the TO220 diodes...
As for bypassing your rectifier... , you need 4 22nf caps, 2 connects to each AC connection point with the oposite leads connecting to + and -
scanned through a few posts.. dont forget there are insulators you can use in the PC PSU on the heatsinks... just unscrew the TO220 diodes...
As for bypassing your rectifier... , you need 4 22nf caps, 2 connects to each AC connection point with the oposite leads connecting to + and -
It doesn't have them. It uses bridges and normal diodes. Anyways, if I have to get TO-220 regulators, I have to get them at Mouser and they have mica there.
That's what I did with 0.0047µFs. The site where I read to do this is telling to use 0.01µF to 0.05µFs
Ok those caps are fine then... everything I read up to now, indicates there is little to gain going any higher than. 0022uf...that little bit wont hurt though...
I am just trying to get a better picture... are you running the amp and Tube off the same transformer?
I am just trying to get a better picture... are you running the amp and Tube off the same transformer?
Yes, same transformers, same rectifiers.
There are two transformers in the amp.
They both have 3 rails :
0-13V
0-13V
0-7V
The four 13VAC 2amp rails are in series and provide the main ±40VDC used for everything. The 7V 0.5amp rails are in parallel and feed the filament regulator.
Before, I was running the 13VAC rails in a 13-0-13 configuration and the tube distorted a little too much due to the ±20VDC supply. There was less hum and almost no wobble.
I figured that even standard LM3x7 regulators will work, no need for rare HV versions. (Because my output voltage is lower than 37V)
Don't worry about the voltage too much with those regulators... what you need to look at is the graph on the datasheet, that shows, how much current they can handle for a specified voltage drop... up to about 40V drop, at which point it can only handle a few mA...
when their is current demand from the chip, it will cause a dip / ripple in the power supply...feeding the tube...
The problem is that the tube realy needs a stable voltage supply, as it is generaly a balanceing act...
Try something else... regulate one 13V tapping down to 6.3V
power the chip useing the other transformer... (please only use 8 or more omhs with this), It should just cover you, if you don't turn the volume up all the way...
going from 40 to 6.3v is about a 34V drop, the regulator is going to have a tough time keeping up with provideing the constant heater current of almost 400mA and will need a very good heatsink - do I recall you complained about the heat before?
Ag damn, now I see the 7V tapping for the fillament... so no worry then about that.
Can I suggest you look for some very basic tutorial on google about loadlines or tube curves... you will see that the 20V is not what caused the massive DC offset you see (speakers wobbleing...)... and willl learn to adjust the bias and operational points in no time...
In fact... look for a nice Soha Design headphone amp, it will show many of the principles you are struggleing with now...I think however we have narrowed it down to a PSU issue... which still probably requires a bias adjustment to the voltages at hand...
This wobbleing will be there if the signal does not pass through a capacitor to stop the DC path...
when their is current demand from the chip, it will cause a dip / ripple in the power supply...feeding the tube...
The problem is that the tube realy needs a stable voltage supply, as it is generaly a balanceing act...
Try something else... regulate one 13V tapping down to 6.3V
power the chip useing the other transformer... (please only use 8 or more omhs with this), It should just cover you, if you don't turn the volume up all the way...
going from 40 to 6.3v is about a 34V drop, the regulator is going to have a tough time keeping up with provideing the constant heater current of almost 400mA and will need a very good heatsink - do I recall you complained about the heat before?
Ag damn, now I see the 7V tapping for the fillament... so no worry then about that.
Can I suggest you look for some very basic tutorial on google about loadlines or tube curves... you will see that the 20V is not what caused the massive DC offset you see (speakers wobbleing...)... and willl learn to adjust the bias and operational points in no time...
In fact... look for a nice Soha Design headphone amp, it will show many of the principles you are struggleing with now...I think however we have narrowed it down to a PSU issue... which still probably requires a bias adjustment to the voltages at hand...
This wobbleing will be there if the signal does not pass through a capacitor to stop the DC path...
Heh, well I saw what you mean. The input is 40V and the output will probably be 35V so I don't have to worry too much. (I-O differential is only 5V)
1. They have a negative voltage filament PSU since they use a dual rail PSU only. (My 6.3v line has separate filaments and the rails are unreferenced.)
2. I see 100µF polarized 'lytics in series with the AC lines.
3. They get the signal from the plate rather than the cathode(But here we invert the signal?)
4. The plate always has some Constant Current Source circuit between it and the positive rail(But there's no voltage regulation). Here, I have NOTHING.
5. There's a pot on the negative rail to adjust bias, and there's a capacitor bypassing them.(Not a good idea in my case as it produces MUCH more hum IIRC.)
I don't know if I said this, but I'd need to have the amp at least working in stereo for Thursday next week(Or I'll be stuck 4 days w/o music).
I'd like to know if my temporary solutions would work or not :
-If there's still hum, remove tube buffer, disconnect 10kOhm and connect the tube input directly to the output.
-If there's no hum but wobble : Put my two 200v 1000µF caps in series with the speakers(It solves the problem and I hear no major sound difference).
Ok, then you have given te answer allready...
If you use the caps you get no wobble (as you call it)
This wobble is absolutley normal, but bad news for anything after it...
However it should only occur at power on and power down if you have it properly biased
Tried referenceing heater to ground yet? simple connection from middle pin to ground should do it...
If you use the caps you get no wobble (as you call it)
This wobble is absolutley normal, but bad news for anything after it...
However it should only occur at power on and power down if you have it properly biased
Tried referenceing heater to ground yet? simple connection from middle pin to ground should do it...
Oh, it's maybe a bias problem then?
Is it poorly adjusted? (+0.86v at tube output / cathode, if that's where I'm supposed to measure it...)
Or is it just moving too much?
Is it poorly adjusted? (+0.86v at tube output / cathode, if that's where I'm supposed to measure it...)
Or is it just moving too much?
A while ago, before I had wobble problems, to try to fix the hum.
No difference...
About the bias, what's the thing, is it moving because of the power amps lowering the rails voltage and changing the bias voltage or is it poorly adjusted and I'm shortening the $50 tube's life?
- Status
- Not open for further replies.