I success design class H amplifier with 120VDC power supply (Schematic from me and layout PCB by my friend). My friend ask me to design class H amplifier with 180VDC and use 3 step of stepper.
Can you suggest me a VAS (Voltage Amplification Stage) transistor which have max. VCEO 400V and low Cob?
Can you suggest me a VAS (Voltage Amplification Stage) transistor which have max. VCEO 400V and low Cob?
Can you suggest me a VAS (Voltage Amplification Stage) transistor
which have max. VCEO 400V and low Cob?
NPN - ZXTN08400BFF is 10pF
PNP - ZXTP08400BFF is 20pF
Last edited:
NPN - ZXTN08400BFF is 10pF
PNP - ZXTP08400BFF is 20pF
Thank you. But we do not use SMD part.
It can TO-126 or TO-220 package.
That is quite hard to find. Crest Audio Used MJE13007/MJE5852 in their high power amplifier.
Currently there is no VAS in the PROx200 series 🙂
Sajti
Currently there is no VAS in the PROx200 series 🙂
Sajti
Last edited:
Reconsider not using SMD parts. You will have a tough time trying to find suitable parts in TO-126 now. Use a cascode arrangement for best results.
You’re going to be stuck either using a cascode or a darlington. Won’t find anything with the low Cob typically needed for high performance AND enough SOA to handle the currents you need (peaks of 20-ish mA). C2752/A1156 are suitable as the darlington *drivers* for MJE18004/MJE5852. A little wimpy for a VAS stand alone. The MJE’s are too big.
Creat did use the MJE’s as VAS in their earlier models. They have abandoned the conventional Lin topology in their Pro 200 series for something that looks Ike the QSC PLX. No transistor ever sees the full 320 volt rail to rail voltage swing - the penalty being an output stage with gain. That has a bit more commutation distortiion than a unity gain output stage, but perhaps more reliable. And relieves the conundrum of finding a VAS transistor with 2 pF Cob, 400 Vceo, Hfe of 200, and 50 mA of SOA at 300V.
Creat did use the MJE’s as VAS in their earlier models. They have abandoned the conventional Lin topology in their Pro 200 series for something that looks Ike the QSC PLX. No transistor ever sees the full 320 volt rail to rail voltage swing - the penalty being an output stage with gain. That has a bit more commutation distortiion than a unity gain output stage, but perhaps more reliable. And relieves the conundrum of finding a VAS transistor with 2 pF Cob, 400 Vceo, Hfe of 200, and 50 mA of SOA at 300V.
Won’t find anything with the low Cob typically needed for high performance AND enough SOA to handle the currents you need (peaks of 20-ish mA).
Performance is not question with class H, any more...
Sajti
You’re going to be stuck either using a cascode or a darlington. Won’t find anything with the low Cob typically needed for high performance AND enough SOA to handle the currents you need (peaks of 20-ish mA). C2752/A1156 are suitable as the darlington *drivers* for MJE18004/MJE5852. A little wimpy for a VAS stand alone. The MJE’s are too big.
Creat did use the MJE’s as VAS in their earlier models. They have abandoned the conventional Lin topology in their Pro 200 series for something that looks Ike the QSC PLX. No transistor ever sees the full 320 volt rail to rail voltage swing - the penalty being an output stage with gain. That has a bit more commutation distortiion than a unity gain output stage, but perhaps more reliable. And relieves the conundrum of finding a VAS transistor with 2 pF Cob, 400 Vceo, Hfe of 200, and 50 mA of SOA at 300V.
Thank you for your suggestion. I will try to sim output as CFP and followed by emitter follower with gain. VAS use medium voltage.
Cascode VAS works well with high current BJTs. I simulated my amplifier with 2SC3503/2SA1381, and with MJE15032/MJE15033. Almost no difference in distortion...
Sajti
Sajti
Search Mackie amplifier schematic class H ( Mackie M2600 + Mackie M3000 + Mackie M4000 ) .
This schematics exceeds the forum's limit of 976.6 KB for attachment ...
This schematics exceeds the forum's limit of 976.6 KB for attachment ...
Attachments
Last edited:
Thank you for your suggestion. I will try to sim output as CFP and followed by emitter follower with gain. VAS use medium voltage.
That would be one method. The Crest (and PLX) simply use commn emitter output stages. They also use op amp input sages, giving them a very high negative feedback factor, which helps keep commutation distortion down to inaudible levels with normal music signals.
The Pro9200 does lend itself to DIY quite well if you just use a conventional toroid power supply. The SMPS isn’t very DIY-friendly, but the amp is. You could also marry the amp section with any number of rail switch varieties - including more than 2 steps if you were so inclined.
Cascode VAS works well with high current BJTs. I simulated my amplifier with 2SC3503/2SA1381, and with MJE15032/MJE15033. Almost no difference in distortion...
Sajti
Only problem with a cascode VAS is that it robs voltage swing. I just hate wasting 3.9 volts if I paid for them with iron and aluminum. If the higher performance matters more than the extra dissipation, go for it.
Only problem with a cascode VAS is that it robs voltage swing. I just hate wasting 3.9 volts if I paid for them with iron and aluminum. If the higher performance matters more than the extra dissipation, go for it.
Agree. In audio pro, max. power is matter than sound quality. But in most case, I will choose quality over max. power.
Only problem with a cascode VAS is that it robs voltage swing. I just hate wasting 3.9 volts if I paid for them with iron and aluminum. If the higher performance matters more than the extra dissipation, go for it.
I use the Marantz verision. It use LED for the cascode, which means only 2,5V loss. Compare it to the 180V psu. There are many solutions to keep the VAS rail higher...
Sajti
Agree. In audio pro, max. power is matter than sound quality. But in most case, I will choose quality over max. power.
You can see my VAS on the picture. Yes, this is not pro audio version, but not a big loss, if you use it. You will loose much more with the triple emitter follower output.
The common emitter output stage of Crest or even the QSC has much better efficiency...
Sajti
Attachments
You can see my VAS on the picture. Yes, this is not pro audio version, but not a big loss, if you use it. You will loose much more with the triple emitter follower output.
The common emitter output stage of Crest or even the QSC has much better efficiency...
Sajti
I often use that cascode (Hawksford cascode), too.
is there another scheme that is almost the same as this scheme?Search Mackie amplifier schematic class H ( Mackie M2600 + Mackie M3000 + Mackie M4000 ) .
This schematics exceeds the forum's limit of 976.6 KB for attachment ...
is there another scheme that is almost the same as this scheme?
Attachments
KSA1381/KSC3503 are both 300V, 1.8pF and TO126. That would be about the limit; a 400V transistor would be rare, since video drivers have gone to the dodo.
In your cascode you can go down to 2.5V loss of headroom. That is acceptable.......
In your cascode you can go down to 2.5V loss of headroom. That is acceptable.......
- Home
- Amplifiers
- Solid State
- VAS for class H amplifier