The ccs came from G pimm's latest self-biasing circuit (http://home.pacifier.com/~gpimm/self_bias.htm).
The shunt regulator is from raleighaudio (http://www.raleighaudio.com/ccs.htm)
I tested the CCS circuit first by placing a 2kohm load from CCS output to ground. Using the trimpot, I have set the output to 75mA. At this time, the shunt regulator is still not connected.
R1 and R2 of the regulator circuit is connected. but the TL431 and the tube are not yet connected in the circuit. I measured around 147V across R1 and around 2.5V across R2. This are the correct measurements.
I then inserted the TL431 and 6SL7GT to the shunt regulator circuit. Voltage across R1 is pretty much the same. But the voltage across R2 has been halved. It dropped from around 2.5V to 1.25V. I thought the reference pin of the TL431 is suppose to keep voltage across R2 at 2.5V. I also measured the voltage on the cathode pin of the TL431 and it is around 1.9V.
any ideas?
========
also when testing regulated B+, do I need a load? With both the CCS and shunt regulator circuits connected and no load, the voltage from triode plate to ground is around 340V. I thought that with regulator power supplies, no load is needed to set the required voltage/current.
My regulated PSU knowledge is limited to LM317 (and other similar chips) where no load is required to set current/voltage.
Thanks for the help.
The shunt regulator is from raleighaudio (http://www.raleighaudio.com/ccs.htm)
I tested the CCS circuit first by placing a 2kohm load from CCS output to ground. Using the trimpot, I have set the output to 75mA. At this time, the shunt regulator is still not connected.
R1 and R2 of the regulator circuit is connected. but the TL431 and the tube are not yet connected in the circuit. I measured around 147V across R1 and around 2.5V across R2. This are the correct measurements.
I then inserted the TL431 and 6SL7GT to the shunt regulator circuit. Voltage across R1 is pretty much the same. But the voltage across R2 has been halved. It dropped from around 2.5V to 1.25V. I thought the reference pin of the TL431 is suppose to keep voltage across R2 at 2.5V. I also measured the voltage on the cathode pin of the TL431 and it is around 1.9V.
any ideas?
========
also when testing regulated B+, do I need a load? With both the CCS and shunt regulator circuits connected and no load, the voltage from triode plate to ground is around 340V. I thought that with regulator power supplies, no load is needed to set the required voltage/current.
My regulated PSU knowledge is limited to LM317 (and other similar chips) where no load is required to set current/voltage.
Thanks for the help.
jarthel:
It sounds like your R1 and R2 are too low. When current is drawn by the tube and reference diode there is less for R1/R2, so voltage across R2 goes down.
Of the total 75mA provided by the source some must go to the load, some through the tube (shunt regulating) and some through the R1/R2 voltage divider. Plan on a few milliamps for R1 and R2, leaving the rest for the load and the shunt. I would try 500 ohms for R2 and 30K for R1. ( I am guessing based on the voltages you mention you are aiming for 150V output, about 70mA ?)
Make sure you do use a load when you set it up since the 6SL7 isnt rated for very much current, so the load must draw most of it. Either that or use a different tube.
It sounds like your R1 and R2 are too low. When current is drawn by the tube and reference diode there is less for R1/R2, so voltage across R2 goes down.
Of the total 75mA provided by the source some must go to the load, some through the tube (shunt regulating) and some through the R1/R2 voltage divider. Plan on a few milliamps for R1 and R2, leaving the rest for the load and the shunt. I would try 500 ohms for R2 and 30K for R1. ( I am guessing based on the voltages you mention you are aiming for 150V output, about 70mA ?)
Make sure you do use a load when you set it up since the 6SL7 isnt rated for very much current, so the load must draw most of it. Either that or use a different tube.
here's the combined schematic:
http://img470.imageshack.us/my.php?image=untitled3vd5.png
thanks for the help
http://img470.imageshack.us/my.php?image=untitled3vd5.png
thanks for the help
Hi Jarthel,
I noted the TL431 reference input bias current is only about 4uA so 100uA current in the reference leg is not unreasonable, although I would probably go for 1mA or so..
Are you certain you have not swapped any of the TL431 pins (gnd and output?)
Also the current handling capability of the 6SL7 is extremely limited, a couple of say 6BX7/6V6/6CK4/12B4 might be a better choice. Under no load conditions all of the load current must flow through the shunt regulator tube. Any combination that can dissipate >12W ought to be ok.
Given that you used a 6SL7 it is highly likely that open circuit it would saturate and kill the TL431.
I noted the TL431 reference input bias current is only about 4uA so 100uA current in the reference leg is not unreasonable, although I would probably go for 1mA or so..
Are you certain you have not swapped any of the TL431 pins (gnd and output?)
Also the current handling capability of the 6SL7 is extremely limited, a couple of say 6BX7/6V6/6CK4/12B4 might be a better choice. Under no load conditions all of the load current must flow through the shunt regulator tube. Any combination that can dissipate >12W ought to be ok.
Given that you used a 6SL7 it is highly likely that open circuit it would saturate and kill the TL431.
kevinkr said:
do you mean I need to decrease R1/R2 combo?
kevinkr said:
I've checked the proper pins several times by comparing the actual chip to the diagram in the datasheet. I do not think I made a mistake in this regard.
kevinkr said:
Where in the datasheet can I find the 12W dissipation info? I am looking at the 6BX7GT datasheet (http://www.mif.pg.gda.pl/homepages/frank/sheets/093/6/6BX7GT.pdf) and cannot seem to find the relevant info.
when using a 6v6, I assumed you meant it's configured as a triode? if yes, any chance of explaining how this is done?
kevinkr said:
just wondering how do you know that a 6sl7 is unsuitable.
Thank you for the help.
Hi Jarthel,
The 6SL7 is an inappropriate choice because:
a)If the supply is unloaded the tube will be called upon to sink the full 75mA, which is about 20 times higher than the tube design target value.
b.) Dissipation per section is 1W for a total of 2W which means that 150V the maximum current the tube can handle before the dissipation limit is exceeded is about 13mA.
Yes, reduce R1/R2 by a factor of 10 to get 1mA of current in the reference resistor string. Say 2.4K and 150K.
6V6 can easily be triode strapped, connect screen grid to plate using a 100 ohm 1/2W resistor.
What is the overall load current this supply is meant to deliver to the load? To assure good regulation under all conditions particularly if the load is not purely class A I would recommend at least 10 - 20 % of the load current in normal operation be diverted to the shunt regulator with the rest going to the load. This would mean for an external load of 75mA max you would want the current source to supply at least an additional 10mA which is "burned up" in the shunt regulator. Note that the tube must be able to dissipate the full power if the load is disconnected for some reason. (Wiring error, tube failure, failed component.)
I used lower power shunt regulators in some of my early pre-amp designs and found them particularly worthwhile for avoiding low frequency interactions through the supplies. (I used one for each gain block.)
Hope this all helps.
The 6SL7 is an inappropriate choice because:
a)If the supply is unloaded the tube will be called upon to sink the full 75mA, which is about 20 times higher than the tube design target value.
b.) Dissipation per section is 1W for a total of 2W which means that 150V the maximum current the tube can handle before the dissipation limit is exceeded is about 13mA.
Yes, reduce R1/R2 by a factor of 10 to get 1mA of current in the reference resistor string. Say 2.4K and 150K.
6V6 can easily be triode strapped, connect screen grid to plate using a 100 ohm 1/2W resistor.
What is the overall load current this supply is meant to deliver to the load? To assure good regulation under all conditions particularly if the load is not purely class A I would recommend at least 10 - 20 % of the load current in normal operation be diverted to the shunt regulator with the rest going to the load. This would mean for an external load of 75mA max you would want the current source to supply at least an additional 10mA which is "burned up" in the shunt regulator. Note that the tube must be able to dissipate the full power if the load is disconnected for some reason. (Wiring error, tube failure, failed component.)
I used lower power shunt regulators in some of my early pre-amp designs and found them particularly worthwhile for avoiding low frequency interactions through the supplies. (I used one for each gain block.)
Hope this all helps.
kevinkr said:Hi Jarthel,
The 6SL7 is an inappropriate choice because:
a)If the supply is unloaded the tube will be called upon to sink the full 75mA, which is about 20 times higher than the tube design target value.
b.) Dissipation per section is 1W for a total of 2W which means that 150V the maximum current the tube can handle before the dissipation limit is exceeded is about 13mA.
Yes, reduce R1/R2 by a factor of 10 to get 1mA of current in the reference resistor string. Say 2.4K and 150K.
6V6 can easily be triode strapped, connect screen grid to plate using a 100 ohm 1/2W resistor.
What is the overall load current this supply is meant to deliver to the load? To assure good regulation under all conditions particularly if the load is not purely class A I would recommend at least 10 - 20 % of the load current in normal operation be diverted to the shunt regulator with the rest going to the load. This would mean for an external load of 75mA max you would want the current source to supply at least an additional 10mA which is "burned up" in the shunt regulator. Note that the tube must be able to dissipate the full power if the load is disconnected for some reason. (Wiring error, tube failure, failed component.)
I used lower power shunt regulators in some of my early pre-amp designs and found them particularly worthwhile for avoiding low frequency interactions through the supplies. (I used one for each gain block.)
Hope this all helps.
is dissipation the only factor for tube selection?
No, dissipation isn't the only factor, things like the continuous and peak cathode/plate currents also play a large role, low rp (for good ac performance) and high transconductance/high perveance should also be beneficial in this application.
The EL34 triode connected would seem to be almost be ideal for this application. Check the curves and see how much bias voltage is required at the target plate current and voltage. (Vp 150V, Ip 10mA - 75mA, effective bias must be less than -30V across the entire range for this to work with the TL431, and I suspect it will be.)
Good luck!
The EL34 triode connected would seem to be almost be ideal for this application. Check the curves and see how much bias voltage is required at the target plate current and voltage. (Vp 150V, Ip 10mA - 75mA, effective bias must be less than -30V across the entire range for this to work with the TL431, and I suspect it will be.)
Good luck!
kevinkr said:
The EL34 triode connected would seem to be almost be ideal for this application. Check the curves and see how much bias voltage is required at the target plate current and voltage. (Vp 150V, Ip 10mA - 75mA, effective bias must be less than -30V across the entire range for this to work with the TL431, and I suspect it will be.)
Good luck!
http://www.mif.pg.gda.pl/homepages/frank/sheets/030/e/EL34.pdf
any chance of pointing me to the right graph? thanks again
If you bias the EL34 at -30V and put 75mA through the TL431 (the no load condition) it is going to FRY.
You need to insert a voltage dropping resistor from the tube cathode such that the voltage seen by the TL431 x Imax(75mA) is < Pd max for the TL431 (check the datasheet for Pd max). This is what Supratek did in their 5881 based preamp shunt reg.
In my opinion a 10 - 75mA current range is too much for a TL431 based shunt reg.
pm
You need to insert a voltage dropping resistor from the tube cathode such that the voltage seen by the TL431 x Imax(75mA) is < Pd max for the TL431 (check the datasheet for Pd max). This is what Supratek did in their 5881 based preamp shunt reg.
In my opinion a 10 - 75mA current range is too much for a TL431 based shunt reg.
pm
mach1 said:
Good point, I wasn't thinking about the dissipation in the TL431.
Actually under the no external load condition the bias will be much less with a vp of 150V, however your point stands.
A resistor would be needed, or perhaps better a npn power transistor/TL431 cascode - this would get around both the dissipation and voltage limitation. (You could still use a tube on top of that or a just one single high voltage transistor) Stability might be questionable though and I haven't tried it. YMMV
Frankly for these current levels this is looking less and less attractive.. This circuit is well suited to applications where the load current is invariant or much lower overall current is required.
Jarthel what is the load current range?
Re the triode connected curve for an EL34
Take a look at p.9 here: http://www.mif.pg.gda.pl/homepages/frank/sheets/129/e/EL34.pdf
This would seem to indicate about 80ma at -5V effective bias and about 40ma at -10V and about 15ma at -15V with an EL34 so although marginal and not recommended it might actually work..
One caveat is that modern EL34 don't necessarily correspond that well to the old mullard curves, I'd download the curve from JJ and take a look at that. (No time left in the day or I would have.)
kevinkr said:
any advice on what transistor to use?
kevinkr said:
68mA
kevinkr said:
This would seem to indicate about 80ma at -5V effective bias and about 40ma at -10V and about 15ma at -15V with an EL34 so although marginal and not recommended it might actually work..
One caveat is that modern EL34 don't necessarily correspond that well to the old mullard curves, I'd download the curve from JJ and take a look at that. (No time left in the day or I would have.)
[/B][/QUOTE]
thank you
Connect the transistor's collector to the cathode of the shunt regulatior tube, connect the base to a voltage source (Could be a resistive voltage divider) of say 4 - 5V and the emitter to the TL431.
Transistor should have a vce of 60V or better, beta >100, hfe 100MHz.
Some details:
The base of the transistor must be set somewhat lower than the minimum bias voltage required for the shunt regulator tube to carry the full no load current.
You can use a resistive divider (bypass cap to ground) or a zener to stabilize the voltage on the base. I'd recommend a zener running at >5mA as base current could be as higher as .68mA depending on transistor beta.
Another alternative might be to use a ground referenced dc filament supply to provide the required base voltage. In this instance use a small series resistor (100 ohms) and a small cap (~100pF) from base to collector to prevent oscillation.
Do you have an oscilloscope? You will need to check for oscillation in this cascode.
Expect to do some tinkering to get this idea to work. Note that this approach can be used to regulate very high voltages! YMMV
An off the cuff analysis of the EL34/TL431 combination (no transistor) on a VP of 150V indicates this would work. Dissipation in the TL431 would be worst case about 400mW (check my math!!) and the dip package with P suffix can dissipate up to 1W at ambient. In normal operation the dissipation should be in the region of 200mW.
68mA load current with a CCS of 75mA seems about ok, perhaps slightly more current would be better, but I think it will work fine.
Transistor should have a vce of 60V or better, beta >100, hfe 100MHz.
Some details:
The base of the transistor must be set somewhat lower than the minimum bias voltage required for the shunt regulator tube to carry the full no load current.
You can use a resistive divider (bypass cap to ground) or a zener to stabilize the voltage on the base. I'd recommend a zener running at >5mA as base current could be as higher as .68mA depending on transistor beta.
Another alternative might be to use a ground referenced dc filament supply to provide the required base voltage. In this instance use a small series resistor (100 ohms) and a small cap (~100pF) from base to collector to prevent oscillation.
Do you have an oscilloscope? You will need to check for oscillation in this cascode.
Expect to do some tinkering to get this idea to work. Note that this approach can be used to regulate very high voltages! YMMV
An off the cuff analysis of the EL34/TL431 combination (no transistor) on a VP of 150V indicates this would work. Dissipation in the TL431 would be worst case about 400mW (check my math!!) and the dip package with P suffix can dissipate up to 1W at ambient. In normal operation the dissipation should be in the region of 200mW.
68mA load current with a CCS of 75mA seems about ok, perhaps slightly more current would be better, but I think it will work fine.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.