I've got:
4 x DN2540
6 x 2N6520
8 x 2N3906
4 x LM317
8 x 2N3904
4 x IRFU310
Seen various designs and discussions involving these and not sure where to begin. Simple cascade involving 2 x DN2540's? So many opinions out there.
Needing a CCS for a plate load on a 6N1P linestage.
Currently using a SV83 as a CCS on the plate but would like to try a solid state CCS. Using led bias on the cathode.
Suggestions would be appreciated.
Thks
4 x DN2540
6 x 2N6520
8 x 2N3906
4 x LM317
8 x 2N3904
4 x IRFU310
Seen various designs and discussions involving these and not sure where to begin. Simple cascade involving 2 x DN2540's? So many opinions out there.
Needing a CCS for a plate load on a 6N1P linestage.
Currently using a SV83 as a CCS on the plate but would like to try a solid state CCS. Using led bias on the cathode.
Suggestions would be appreciated.
Thks
Last edited:
You may need a heat sink on the top FET, since it has most of the voltage across it. But maybe not- it depends on what you're trying to get your circuit to do- if it's a low level circuit with minimal swing requirements (a few volts), then you can set the B+ to be 30V or so higher than the predicted plate voltage. That way, dissipation at 20mA will be half a watt or less. If you're trying to swing lots of volts, then yeah, you'll have to make the B+ higher and heat sink the top one.
I've used that combo. Some folk will advise higher gate stopper values, but I found 220R (1/4 Watt) worked well. Heat sink the upper FET. Look to it's datasheet and work out what sink you need... good to learn it (...after all SY can't hold your hand all year).
Used on a 6H30 at 25mA the performance of the cascode FETs was mighty fine. Should be just as good on your tube. You need a few volts headroom across the FETs (i.e. above your maximum positive swing). I can't recall what it was... about 15V I think.
Used on a 6H30 at 25mA the performance of the cascode FETs was mighty fine. Should be just as good on your tube. You need a few volts headroom across the FETs (i.e. above your maximum positive swing). I can't recall what it was... about 15V I think.
Hey Brit01,
A few points here:
1) Mount the stopper resistors (the 220R) as close to the body of the DN2540 gate as possible. I typically cut the resistor lead to a minimum, about 1/4" or so; then fold it and attach as close as I can to the body of the MOSFET. I did use 1k0 resistors, SEI (Stackpole) from Handmade Electronics.
2) Use carbon comp or carbon film resistors for stoppers. Do not use metal films in this application. No current is flowing, so don't fret about noise.
3) I had good luck using nice 10+ turn Cermet trimpots to set current instead of the fixed resistor. There is enough variation in the MOSFET devices to drive you crazy trying to get the current dead on. No problems with noise or instability that I can tell.
4) This setup is very nice (the cascoded DN2540). It sounds much nicer than the C4S, with LEDs for voltage references or with 2.5V LT1004 types. You'll be quite happy with it.
5) The DN2540 is available in TO92 as well as TO220. For the lower device, a TO92 works just fine (1 watt rated, although I wouldn't want to try it) and is 1/2 the cost.
Have fun,
Stuart
A few points here:
1) Mount the stopper resistors (the 220R) as close to the body of the DN2540 gate as possible. I typically cut the resistor lead to a minimum, about 1/4" or so; then fold it and attach as close as I can to the body of the MOSFET. I did use 1k0 resistors, SEI (Stackpole) from Handmade Electronics.
2) Use carbon comp or carbon film resistors for stoppers. Do not use metal films in this application. No current is flowing, so don't fret about noise.
3) I had good luck using nice 10+ turn Cermet trimpots to set current instead of the fixed resistor. There is enough variation in the MOSFET devices to drive you crazy trying to get the current dead on. No problems with noise or instability that I can tell.
4) This setup is very nice (the cascoded DN2540). It sounds much nicer than the C4S, with LEDs for voltage references or with 2.5V LT1004 types. You'll be quite happy with it.
5) The DN2540 is available in TO92 as well as TO220. For the lower device, a TO92 works just fine (1 watt rated, although I wouldn't want to try it) and is 1/2 the cost.
Have fun,
Stuart
It's not you. The schematic drawing program I had at the time only had P. The devices are actually N channel. Pretend the arrows go the other direction. Or use the correct schematics from the ImPasse article that explains the use of the simple cascodes.
Great information guys. thxs very much. this is the first time I've tried a CCS with FETS. Was previously using a pentode (with fine results but sucks up so much voltage among other complications).
I've got the N3 type of Fet. Do I need the N5 type for this application? You mentioned it will be ok for the lower device. so I need the N5 for the upper? (which I don't have )
I see now TO92 N3 can handle only 1 watt but the N5 15w.
Any way of substituting the upper T092 for another device (I need to look at these further when I get home; at work now) :
6 x 2N6520
8 x 2N3906
8 x 2N3904
4 x IRFU310
I've got the N3 type of Fet. Do I need the N5 type for this application? You mentioned it will be ok for the lower device. so I need the N5 for the upper? (which I don't have )I see now TO92 N3 can handle only 1 watt but the N5 15w.
Any way of substituting the upper T092 for another device (I need to look at these further when I get home; at work now) :
6 x 2N6520
8 x 2N3906
8 x 2N3904
4 x IRFU310
Last edited:
Brit01,
You'll probably be best off not substituting another device in the upper position. The wattage of the device needed depends on what you want your plate current and voltage to be. With a 6N1P, it may be safe to assume you're looking for ~150V on the plate? If so, and with that 220V B+, the 70V drop will almost all be across the upper device. That said, and with a 1W device, I'd probably keep the current to no more than 10mA, so that upper dissipation will be ~0.7 watts. Keep lots of air around the TO92 device. If you suspect high ambient temperatures around the device, cut your dissipation to no more than 0.5W, which means raising plate voltage (not B+) or lowering current.
On a side note, but important here, is the method of bias. Unless you're working with plate curves and load lines, if you're breadboarding, and using resistor bias, a pot will work on a temporary basis. You could simply adjust the pot to find the correct value for the current/plate voltage you want. You'll have to also adjust the CCS pot to obtain the current you want. But, if you will be using solid state bias or battery bias, then you'll need to select from an available voltage (from the battery of a combination of diodes), then adjust your CCS pot to get the plate voltage and current desired.
LEDs sound good, but SiC Schottkies are much better. ~0.84-0.86 V per diode in the typical input tube current ranges, two SiC diodes will substitute for one red LED. Smoother sound, less solid state edge in my experience. I've used the Cree 600V successfully. I think you'll like it with your CCS.
Attached is the schematic from Gary Pimm's website of how to connect the devices, but the "stopper" resistors are omitted.
Stuart
You'll probably be best off not substituting another device in the upper position. The wattage of the device needed depends on what you want your plate current and voltage to be. With a 6N1P, it may be safe to assume you're looking for ~150V on the plate? If so, and with that 220V B+, the 70V drop will almost all be across the upper device. That said, and with a 1W device, I'd probably keep the current to no more than 10mA, so that upper dissipation will be ~0.7 watts. Keep lots of air around the TO92 device. If you suspect high ambient temperatures around the device, cut your dissipation to no more than 0.5W, which means raising plate voltage (not B+) or lowering current.
On a side note, but important here, is the method of bias. Unless you're working with plate curves and load lines, if you're breadboarding, and using resistor bias, a pot will work on a temporary basis. You could simply adjust the pot to find the correct value for the current/plate voltage you want. You'll have to also adjust the CCS pot to obtain the current you want. But, if you will be using solid state bias or battery bias, then you'll need to select from an available voltage (from the battery of a combination of diodes), then adjust your CCS pot to get the plate voltage and current desired.
LEDs sound good, but SiC Schottkies are much better. ~0.84-0.86 V per diode in the typical input tube current ranges, two SiC diodes will substitute for one red LED. Smoother sound, less solid state edge in my experience. I've used the Cree 600V successfully. I think you'll like it with your CCS.
Attached is the schematic from Gary Pimm's website of how to connect the devices, but the "stopper" resistors are omitted.
Stuart
Attachments
Crossing replies here Brit01,
Only the difference between B+ and plate voltage appears across the CCS. With 220V B+, 15mA and 150V on the plate means 1.05W total dissipation. Either RAISE plate voltage, lower B+ or lower your current unless you will be using TO220 devices.
The other choice is to add heat sinking.
Stuart
Only the difference between B+ and plate voltage appears across the CCS. With 220V B+, 15mA and 150V on the plate means 1.05W total dissipation. Either RAISE plate voltage, lower B+ or lower your current unless you will be using TO220 devices.
The other choice is to add heat sinking.
Stuart
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.