Hello snax,
I'd like to thank you for your interest and offerings
No need to worry about the project being dead there, It's just frozen in between working and other hobbies.. I'm still trying various concepts in search of a good way to achieve thermal and electric stability before actually building the first prototype. When the amp gets to the building stage, we could always see what's up then. Personally I like the idea of a nice enclosure, that's one of my weak points lol. I'm marvelled with some of the casings I've seen on this forum
I'd like to thank you for your interest and offerings
No need to worry about the project being dead there, It's just frozen in between working and other hobbies.. I'm still trying various concepts in search of a good way to achieve thermal and electric stability before actually building the first prototype. When the amp gets to the building stage, we could always see what's up then. Personally I like the idea of a nice enclosure, that's one of my weak points lol. I'm marvelled with some of the casings I've seen on this forum Hi guys =)
It's been a while and recently I picked up my amp building hobby again. It's fun to be doing so again and it hasn't been without result either. I've now almost completed a revision of this SOM500 amp and its concept and the end result is so much better! It's much more efficient, clean, stable, more bandwidth, easily scalable, it's crazy. Cool part is that it's still FET only, but no longer as a goal to avoid using BJTs, but now solely because they're the best choice for this topology.
A quick rundown of distortion figures as they are now:
THD20K @ 500W/4Ohm : 110 ~ 150ppb
THD20K @ 1W/4Ohm : 20 ~ 60ppb
At 100KHz the THD and SINAD meter still show 0.000% / 100.00dB
Soon I'll post up sim results and parts of the schematics.
It's been a while and recently I picked up my amp building hobby again. It's fun to be doing so again and it hasn't been without result either. I've now almost completed a revision of this SOM500 amp and its concept and the end result is so much better! It's much more efficient, clean, stable, more bandwidth, easily scalable, it's crazy. Cool part is that it's still FET only, but no longer as a goal to avoid using BJTs, but now solely because they're the best choice for this topology.
A quick rundown of distortion figures as they are now:
THD20K @ 500W/4Ohm : 110 ~ 150ppb
THD20K @ 1W/4Ohm : 20 ~ 60ppb
At 100KHz the THD and SINAD meter still show 0.000% / 100.00dB
Soon I'll post up sim results and parts of the schematics.
Hi, MagicBox
Absolutely - and thank you for the answers
I was asking about device type because I also used BS170 & BS250 with good results.
Maybe one day I'll try BJT vs MOSFET as cascode devices to a JFET (same amp) - just curios subjectively on sound, if any difference. Vds for JFET also around 18-20V.
Regards,
Tibi
Used the BS types in the amp in this thread or some other?
Most likely in another amp of his.. This one hasn't been built yet though It's a new design with which at some point I got stuck because I did not want to go for the easy way out and having to dealing with stuff like temp and other kinds of instability issues. What's an amp with good numbers on paper that behaves like crap in reality?
The revision deals with all of that. SOM500 was the conceptual design. The revision is the practical design with all the kinks worked out - and it looks very promising.
As for the BS170/250 pairs, yes you can use them throughout in the new schematic. I have hundreds of each stashed somewhere around here They have been obsoleted from what I understand so for the simulation and schematic I used the newer Zetex fets.
It's a new design with which at some point I got stuck because I did not want to go for the easy way out and having to dealing with stuff like temp and other kinds of instability issues. What's an amp with good numbers on paper that behaves like crap in reality?The revision deals with all of that. SOM500 was the conceptual design. The revision is the practical design with all the kinks worked out - and it looks very promising.
As for the BS170/250 pairs, yes you can use them throughout in the new schematic. I have hundreds of each stashed somewhere around here
They have been obsoleted from what I understand so for the simulation and schematic I used the newer Zetex fets.
Last edited:
I'm now finalizing the VAS autobias servo, but ultimately I want the sense not from the VAS current but from the output stage quisent current. But as it is right now with the VAS current biassing it works great already, the qiesent current doesn't budge and as a result the output stage bias is rock solid as well.
Mind you it won't be a complete schematic, powersupplies for now won't be documented so I don't think the schematic would get you off building right away, besides, I need to build and validate the design first before I would even dare suggesting others to build it
I'm thinking of starting a new thread for the amp though as this thread isn't of much relevance any more other than history about how the conceptual design came to be.
Mind you it won't be a complete schematic, powersupplies for now won't be documented so I don't think the schematic would get you off building right away, besides, I need to build and validate the design first before I would even dare suggesting others to build it
I'm thinking of starting a new thread for the amp though as this thread isn't of much relevance any more other than history about how the conceptual design came to be.
Sounds good MB. Not that I need anything more than just a schematic of the amp to try it out. PSU, PCB and everything else are non-issues.
Is the bias rock steady on sims or in actual circuit and at what drive levels? Was the previous version of the SOM500 built by anyone, just for comparison?
Is the bias rock steady on sims or in actual circuit and at what drive levels? Was the previous version of the SOM500 built by anyone, just for comparison?
It's rock steady in the sim, but it should be ust as much for real too; the output stage won't suffer thermal runaway and most schematic's VAS biases tend to shift a little too due to temperature drift. None of this will happen because it's a bias servo.
SOM500 hasn't been built by anyone and hopefully won't be as that's an old conceptual version. I plan on building the revision which I have renamed MF-500 when I complete the amp completely. If you are really interested in buliding one I'd recommend to let me build a prototype first. While from experience with my MF-80 design my sim results are very close to the real thing the real thing may still need some minor tweaking which I will perform on the tech bench No I'm not going to send of DIY's with a sim schematic that hasn't been tested yet, I can't be responsible for the results untill I finished the amp
SOM500 hasn't been built by anyone and hopefully won't be as that's an old conceptual version. I plan on building the revision which I have renamed MF-500 when I complete the amp completely. If you are really interested in buliding one I'd recommend to let me build a prototype first. While from experience with my MF-80 design my sim results are very close to the real thing the real thing may still need some minor tweaking which I will perform on the tech bench
No I'm not going to send of DIY's with a sim schematic that hasn't been tested yet, I can't be responsible for the results untill I finished the amp
Last edited:
Thank you for your interest, I can't wait myself to get to building but first I gotta make sure everything is worked out well before building to avoid having to do multiple PCB runs just because of an oversight. I have my PCBs made so it costs both time and money.
The "VAS" current bias servo has been completed and it works awesome - it does not negatively impact THD numbers at any frequency range. Secondly I am not going to attempt to bias the output bias current as that is the actual signal current and as such will cause drift with the bias setting, something I do not want. On the contrary, the VAS current is a steady-state current when the amp is in normal operating range and as such is completely decoupled from a supposed moving DC in the signal and as such, will always affix the VAS bias to the set current irregardless of signal type / load. In hindsight, this is perfect behaviour.
There's one last thing I need to do now and that is to convert the few ideal current sources in the schematic to real implementations, none of them will noticably impact THD. When that is done, I'll expose the schematic in a new design/build thread
The "VAS" current bias servo has been completed and it works awesome - it does not negatively impact THD numbers at any frequency range. Secondly I am not going to attempt to bias the output bias current as that is the actual signal current and as such will cause drift with the bias setting, something I do not want. On the contrary, the VAS current is a steady-state current when the amp is in normal operating range and as such is completely decoupled from a supposed moving DC in the signal and as such, will always affix the VAS bias to the set current irregardless of signal type / load. In hindsight, this is perfect behaviour.
There's one last thing I need to do now and that is to convert the few ideal current sources in the schematic to real implementations, none of them will noticably impact THD. When that is done, I'll expose the schematic in a new design/build thread
You curious man. This is one of the issues that led me to radically redesign the stage that's typically called "VAS". It's basically two stages; the VCGS (Voltage controlled Current Gain Stage) and CVCS (Current controlled Voltage Converter Stage). These two stages are so intertwined it might as well be called one stage.
This "VAS" does not 'choke' on supply rails and can not be saturated before the amp gets into the instability area (too low Cdom). Thus to answer your question, the VAS itself will never clip under normal conditions and does not suffer a high degree of dislinearity (Vbe slope as is common with powerrail affixed VAS devices). The VCGS works in full class A modus.
In fact, the amp has been designed such way that all loop control signals (Input LTP output, VAS input etc) have optimal swing margin upwards to a number of volts and still keep the amp going without saturating or overdriving anything. However, we're talking 200KHz+ F @ 500W, you won't torture your amp and speakers with such signal in your livingroom, at least, not me
Last edited:
I suggest you call it "The 24-Valve Engine"!
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.