Hello guys - I already use various CCS on my input stages taken from Morgan Jones' book. I now want to try a CCS on the output stage. What I want to use is:
1) a pair of 6AH4 in PP
2) two pairs of 6AH4 in PPP
I'm guessing I need a CCS for each pair - this was recommended by Allen Wright for instance. Requirements are up to 30mA per 6AH4. If this works well I'll try it on 6CK4, 12b4, 6S4A and 1626 as well as 6V6 in triode, so could be up to 40mA for some of those.
So what should I use? Any Schematics or guides I can use to construct one? Many thanks for your opinions - Andy Evans
1) a pair of 6AH4 in PP
2) two pairs of 6AH4 in PPP
I'm guessing I need a CCS for each pair - this was recommended by Allen Wright for instance. Requirements are up to 30mA per 6AH4. If this works well I'll try it on 6CK4, 12b4, 6S4A and 1626 as well as 6V6 in triode, so could be up to 40mA for some of those.
So what should I use? Any Schematics or guides I can use to construct one? Many thanks for your opinions - Andy Evans
If you aren't opposed to sand try the simplest CCS I've ever seen. Use an LM317 voltage regulator, between the ADJ pin and the OUT pin there will always be 1.25 volts. Place an appropriately sized resistor between ADJ and OUT. Bingo, you are done.
In your case you want 60mA so a 20 ohm resistor will give you 62.5mA. Probably close enough for government work.
In your case you want 60mA so a 20 ohm resistor will give you 62.5mA. Probably close enough for government work.
I use the IXYS 10M45 integrated circuit. I have used them in preamp stages, and output stages. To use any CCS in an output stage you will need to increase the supply voltage considerably to provide headroom for the output to swing above the original supply voltage. This is provided by the output transformers inductance in a conventional design. I have a SE 45 amp using the IXYS chip and a supply voltage of 500 volts. Quiescent plate voltage is 275 volts. The limiting factor is the chips maximum voltage rating of 450 volts. Max current is 100mA and max dissipation is 40 watts, and a good heat sink is required. IXYS lists a 900 volt version on their web site but they are not available.
Gary Pimms web site has several CCS designs, but I haven't tried any of them.
http://home.pacifier.com/~gpimm/
There is another thread here where several of us have discussed some simpler hybrid solutions, but I don't think anyone has built anything yet. I am planning a really big hybrid CCS that can power an SE 845. I will probably use an 813, just bought some on Ebay, haven't got them yet.
http://www.diyaudio.com/forums/showthread.php?s=&threadid=67437
Gary Pimms web site has several CCS designs, but I haven't tried any of them.
http://home.pacifier.com/~gpimm/
There is another thread here where several of us have discussed some simpler hybrid solutions, but I don't think anyone has built anything yet. I am planning a really big hybrid CCS that can power an SE 845. I will probably use an 813, just bought some on Ebay, haven't got them yet.
http://www.diyaudio.com/forums/showthread.php?s=&threadid=67437
I have seem circuits that use a CCS for each pair, and circuits that use a CCS for each tube. I haven't tried either so I can't comment on the merits of either design.
The statement I made about increasing the supply voltage only applies if you are using the CCS as the plate load (chokeless parafeed).
I get my IXYS chips from DigiKey in the US. I believe that they ship internationally, but it may cost a lot for a small order. They also carry some good Panasonic electrolytics in 500 volt sizes, and many types of metal film resistors.
http://www.digikey.com/
I went to their web site, click on the UK flag at the top. The banner says "free shipping on orders over 100 pounds sterling". Put 10M45 in the search box and hit enter. Then choose the IC.
The statement I made about increasing the supply voltage only applies if you are using the CCS as the plate load (chokeless parafeed).
I get my IXYS chips from DigiKey in the US. I believe that they ship internationally, but it may cost a lot for a small order. They also carry some good Panasonic electrolytics in 500 volt sizes, and many types of metal film resistors.
http://www.digikey.com/
I went to their web site, click on the UK flag at the top. The banner says "free shipping on orders over 100 pounds sterling". Put 10M45 in the search box and hit enter. Then choose the IC.
CCS's for output stage
Andy,
It is a fact that the CCS for an output stage does not have to be quite as good (as high an AC impedance) as one used on an input diff amp.
I have an EL84 Ultarlinear Class AB1 PP Amp in which I use "Ring of Two" transistor current sources in each tube cathode as shown in the morgan Jones book. I run these at 40mA (using BC547B and MJE340 for the bottom and top transistors respectively). I feel confident that this arrangemet could handle 60mA for a Class A push pull differential output pair.
When it comes to the 120mA for a PPP arrangement of higher current you may wish to try this simple circuit which is for a 300mA CCS. Its not a GREAT CCS (like the Gary Pimm jobs and similar circuits) but its good enough for the application. Caution - max of 36 volts across the TL431. If necessary use some fixed bias to the grids as well. To change its current change that 8R2 resistor - you want 2.5 volts across that resistor so for 120mA CCS use 20.83 Ohms, 47R parallel 39R will be close enough. That "Bias" label means to a positive supply such that there is 1mA (minimum) flowing through the TL431
Cheers,
Ian
Andy,
It is a fact that the CCS for an output stage does not have to be quite as good (as high an AC impedance) as one used on an input diff amp.
I have an EL84 Ultarlinear Class AB1 PP Amp in which I use "Ring of Two" transistor current sources in each tube cathode as shown in the morgan Jones book. I run these at 40mA (using BC547B and MJE340 for the bottom and top transistors respectively). I feel confident that this arrangemet could handle 60mA for a Class A push pull differential output pair.
When it comes to the 120mA for a PPP arrangement of higher current you may wish to try this simple circuit which is for a 300mA CCS. Its not a GREAT CCS (like the Gary Pimm jobs and similar circuits) but its good enough for the application. Caution - max of 36 volts across the TL431. If necessary use some fixed bias to the grids as well. To change its current change that 8R2 resistor - you want 2.5 volts across that resistor so for 120mA CCS use 20.83 Ohms, 47R parallel 39R will be close enough. That "Bias" label means to a positive supply such that there is 1mA (minimum) flowing through the TL431
Cheers,
Ian
Attachments
SY said:
I am definitely not an expert and I've only personally used an LM-317 as a CCS once. I have a single-ended EL84 stereo amp and decided to do this to one channel as a test. I do not have "golden ears" (perhaps they would be more accurately described as "iron"
) but I could hear a definite positive improvement in that channel. Describing the change in words is difficult though. Maybe "clearer" or "more transparent"?However this isn't exactly a high-power amp and being single-ended I think it would be pretty unlikely to overload the 317 so it probably isn't the best test. I've not done any testing other than listening at this point.
I remember that circuit. Input pin of the 7805 goes to the cathode of the tube, ground pin is grounded, and a fixed resistor, or pot goes from the output pin to ground.
The resistor value is chosen to draw the desired current from 5 volts. A bypass cap is needed from cathode to ground to prevent oscillation. The fixed resistor is fed by a regulated 5 volt supply, this causes a constant current to flow, the input pin will float in the range of 7 to 35 volts to force the desired current. In practice this should be used with tobes that want 10 to 30 volts of bias.
I tried this a few years ago in search of an auto bias (selectable) circuit for a SE guitar amp. There is definitely a strange distortion when pounded into hard clipping that doesn't sound good at all. The distortion did not occur with a resistor, so I wound up using a switch and 3 different resistors. This idea might work OK in a HiFi application where you don't purposely drive it into hard clipping on a daily basis.
The resistor value is chosen to draw the desired current from 5 volts. A bypass cap is needed from cathode to ground to prevent oscillation. The fixed resistor is fed by a regulated 5 volt supply, this causes a constant current to flow, the input pin will float in the range of 7 to 35 volts to force the desired current. In practice this should be used with tobes that want 10 to 30 volts of bias.
I tried this a few years ago in search of an auto bias (selectable) circuit for a SE guitar amp. There is definitely a strange distortion when pounded into hard clipping that doesn't sound good at all. The distortion did not occur with a resistor, so I wound up using a switch and 3 different resistors. This idea might work OK in a HiFi application where you don't purposely drive it into hard clipping on a daily basis.
I remember that circuit. Input pin of the 7805 goes to the cathode of the tube, ground pin is grounded, and a fixed resistor, or pot goes from the output pin to ground. >
I know Chris Found, the Beard service engineer, who was involved in designing a lot of the gear, and he suggested I try it. I hated it - there was (??) maybe a bit more clarity, but there was a rough tinny overtone that was quite unbearable - I tore it out.
Whatever goes in the cathode of the output tube must be at least as good as a resistor bypassed by a polypropylene motor run cap of value 100uF. Is this possible? Andy
I know Chris Found, the Beard service engineer, who was involved in designing a lot of the gear, and he suggested I try it. I hated it - there was (??) maybe a bit more clarity, but there was a rough tinny overtone that was quite unbearable - I tore it out.
Whatever goes in the cathode of the output tube must be at least as good as a resistor bypassed by a polypropylene motor run cap of value 100uF. Is this possible? Andy
Hi Andy,
Generally mosfets need more voltage (>20V for a DN2540) across them to bring capacitance down, that implies a heatsink at your aimed 80mA or 160mA.
For that reason I went with bipolars. I use the TIP125, a Darlington, with about 4-5V across it. Havent measured, but it sounds good You could probably go with a cascode of these as well.
Other nice thing of bipolars is that their VBE is constant at 0,6-0.7 volts. That implies that one can use one pot to set the same current at multiple CCS (this is specially interesting when using one CCS per cathode, so 4 CCS for a stereo amp, and connecting the cathodes with a large cap for differential operation). But knowing you, you will use DHTs and avoid caps, so you will end up having to use one CCS per pair of output tubes.
Generally mosfets need more voltage (>20V for a DN2540) across them to bring capacitance down, that implies a heatsink at your aimed 80mA or 160mA.
For that reason I went with bipolars. I use the TIP125, a Darlington, with about 4-5V across it. Havent measured, but it sounds good
You could probably go with a cascode of these as well. Other nice thing of bipolars is that their VBE is constant at 0,6-0.7 volts. That implies that one can use one pot to set the same current at multiple CCS (this is specially interesting when using one CCS per cathode, so 4 CCS for a stereo amp, and connecting the cathodes with a large cap for differential operation). But knowing you
, you will use DHTs and avoid caps, so you will end up having to use one CCS per pair of output tubes.
Hi Andy,
Generally mosfets need more voltage (>20V for a DN2540) across them to bring capacitance down, that implies a heatsink at your aimed 80mA or 160mA.
For that reason I went with bipolars. I use the TIP125, a Darlington, with about 4-5V across it. Havent measured, but it sounds good
Other nice thing of bipolars is that their VBE is constant at 0,6-0.7 volts. That implies that one can use one pot to set the same current at multiple CCS (this is specially interesting when using one CCS per cathode, so 4 CCS for a stereo amp, and connecting the cathodes with a large cap for differential operation). But knowing you
Ah - Erik! Great. I have no real knowledge of solid state so it would be a question of getting a CCS designed by someone who's good with this, and then just building it from a schematic. Do you have one?
Ive used Lm317 with harsh results on the top end. Cascoding as per Walt Jung
http://waltjung.org/PDFs/Sources_101_P2.pdf
makes it more usable. Ive adapted the cascode with enhancement fets with good results. Tl431 is an amazing little device to measure current too. Love using the tiny smd version in cascode with a bigger device. The Lm 317 and its sister are very useful to establish a circuit quickly and easily but expect them to screech if forced to work alone ;0)
TM
http://waltjung.org/PDFs/Sources_101_P2.pdf
makes it more usable. Ive adapted the cascode with enhancement fets with good results. Tl431 is an amazing little device to measure current too. Love using the tiny smd version in cascode with a bigger device. The Lm 317 and its sister are very useful to establish a circuit quickly and easily but expect them to screech if forced to work alone ;0)
TM
When using a 317, you need to confirm that the device is always operating in its regulation range (input to output voltage doesn't dynamically drop below a few volts or peak too high), as the common cathode voltage in a class A PP output stage, triode mode, will have AC ripple and signal voltages on it which can vary markedly with output power level.
The max voltage capability of a 317 CCS can be enhanced by using a resistor in series to drop some of the cathode bias voltage, and reduce thermal loading on the 317, and by putting a zener diode across it.
The max voltage capability of a 317 CCS can be enhanced by using a resistor in series to drop some of the cathode bias voltage, and reduce thermal loading on the 317, and by putting a zener diode across it.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.