Thanks Ilimzn, phew, much information to chew on !
About the 5k in the floating cascode, i considered using a zener instead. According to sims, the floating cascode makes the amp also insensitive to high impedance sources. In simulations distortions only rised ~10% when having a 22k in series with the input.
The wilson current mirror appeared helpful in 2 ways, better balancing and no thermal problems, as the 2 lower devices have nothing/same to dissipate. This should lower dc-offset below 0.1mv. (optimistic
)
The cascoded vas will be the most interesting, i am really curious if the actual existing 2nd harmonic improves the sound or not.
The very low AC-swing in input is IMHO one of the biggest pros of symasym. This in combination with the very balanced operation of inputdiffamp should keep input devices easily in linear operation.
I once tried degeneration (47ohms) in input, but the result was degraded sound. Possible that the large nfb compensate my "lousy" layout. At least that showed me that large nfb is anything but evil.
Cascoding the drivers was just some idea...
Mike
About the 5k in the floating cascode, i considered using a zener instead. According to sims, the floating cascode makes the amp also insensitive to high impedance sources. In simulations distortions only rised ~10% when having a 22k in series with the input.
The wilson current mirror appeared helpful in 2 ways, better balancing and no thermal problems, as the 2 lower devices have nothing/same to dissipate. This should lower dc-offset below 0.1mv. (optimistic
)The cascoded vas will be the most interesting, i am really curious if the actual existing 2nd harmonic improves the sound or not.
The very low AC-swing in input is IMHO one of the biggest pros of symasym. This in combination with the very balanced operation of inputdiffamp should keep input devices easily in linear operation.
I once tried degeneration (47ohms) in input, but the result was degraded sound. Possible that the large nfb compensate my "lousy" layout. At least that showed me that large nfb is anything but evil.
Cascoding the drivers was just some idea...
Mike
OK, thoughts whilst lying in the bath:
This could quite easily be modular. For experimental purposes we could make a series of single layer boards for easy home fabrication, an output board that contains the zobel and big trannies, (up to three pairs for those that require them), and another that has just the basic gain stages and decoupling. This main board would just have a lot of links to the supply rails and various stages, and the assorted cascodes, CCSs etc. would stack up in a 3 dimensional style on top, and could easily be swapped out or altered as required.
Comments?
This could quite easily be modular. For experimental purposes we could make a series of single layer boards for easy home fabrication, an output board that contains the zobel and big trannies, (up to three pairs for those that require them), and another that has just the basic gain stages and decoupling. This main board would just have a lot of links to the supply rails and various stages, and the assorted cascodes, CCSs etc. would stack up in a 3 dimensional style on top, and could easily be swapped out or altered as required.
Comments?
Hi Mike !
Thanks for being so fast !
I didn't see your answer, and I wrote a new post at the symasym topic (that big one) !
Yes, I need 150W, or something little less, because it will be used with 100W RMS speakers in all kinds of ambience, including large ones with pretty much people.
If 150W is so complicated, 125W is easier ? I saw a schematic of Symasym 5 with 125W, but Destroyer couldn't guarantee that it works and I didn't see anybody building.
I'm not expert in SS devices, I don't know what is VAS...
Can I use MJ15024/25 at the output ?
Thanks again Mike.
Regards,
Fernando
Thanks for being so fast !
I didn't see your answer, and I wrote a new post at the symasym topic (that big one) !
Yes, I need 150W, or something little less, because it will be used with 100W RMS speakers in all kinds of ambience, including large ones with pretty much people.
If 150W is so complicated, 125W is easier ? I saw a schematic of Symasym 5 with 125W, but Destroyer couldn't guarantee that it works and I didn't see anybody building.
I'm not expert in SS devices, I don't know what is VAS...
Can I use MJ15024/25 at the output ?
Thanks again Mike.
Regards,
Fernando
Hi Fernando,
I'll give your question a stab.
As the power goes up, so do the voltages and currents required. Heat dissipated is given by the voltage x the current, so very quickly we need to change to heavier devices that do not have the same characteristics as the ones that worked so well. This is true for every part in the amp pretty much.
The Vas is the Voltage Amplification Stage. VAS.
-Chris
I'll give your question a stab.
As the power goes up, so do the voltages and currents required. Heat dissipated is given by the voltage x the current, so very quickly we need to change to heavier devices that do not have the same characteristics as the ones that worked so well. This is true for every part in the amp pretty much.
The Vas is the Voltage Amplification Stage. VAS.
Yes, but the performance may be lower. This means less performance than a normal Symasym, but may be higher than most amplifiers.
-Chris
Hi busy Mike!,
Hi Al,
I see two boards. Front end and current amplification stage. This allows us to run the front end from it's own supply (my intention on a better Symasym). We could then play with different output stage setups.
One big board to test concepts would work as well. This goes far beyond what I think the original concept was.
Mike, Pavel, any feelings on this?
-Chris
That would be the purpose of this schematic and board. So yes!
Yes, a worthwhile baby step.
Hi Al,
I see two boards. Front end and current amplification stage. This allows us to run the front end from it's own supply (my intention on a better Symasym). We could then play with different output stage setups.
One big board to test concepts would work as well. This goes far beyond what I think the original concept was.
Mike, Pavel, any feelings on this?
-Chris
anatech said:
Oh, nice, but the amp efficiency is very important to it's realiability and less power consuption / dissipation, and of course more output power.
I don't want to loose quality to have more power at the output, can I bridge two symasym? Will it degrade it's quality ?
Thank you !
Regards,
Fernando
Fernando, because a speaker is rated 100w you do not need to have 100w amp. It is possible to increase poweroutput for symasym, but you will need to replace all the cheap&nice 2n5401/2n5551 with exotic hard to get expensive devices, and readjust feedback compensation. Chew on that: to double the perceived volume you need ten times the power, 60watt -> 600watt ! Doubling outputpower is barely noticeable.
If you need that much power, there are also super leach amp and the Quasi's.
MJ15024/25 are nice transistors (if original) with an unique feature, they have a constant/flat ft of 7mhz for the first amperes. With an idle bias of ~110ma they should work good. It just might be tough to connect TO3s to the existing PCBs. (The drivers should be connected to the same heatsink)
Chris, Al, i think a single board prepared for all features should do it for the beginning, just to find out what to keep. As long as symasym is fed with 36v, i do not see any advantage in more outputdevices.
For a real high power version it will be necessary to use to126 devices instead of the cute to92, needing a new board anyway.
Mike
If you need that much power, there are also super leach amp and the Quasi's.
MJ15024/25 are nice transistors (if original) with an unique feature, they have a constant/flat ft of 7mhz for the first amperes. With an idle bias of ~110ma they should work good. It just might be tough to connect TO3s to the existing PCBs. (The drivers should be connected to the same heatsink)
Chris, Al, i think a single board prepared for all features should do it for the beginning, just to find out what to keep. As long as symasym is fed with 36v, i do not see any advantage in more outputdevices.
For a real high power version it will be necessary to use to126 devices instead of the cute to92, needing a new board anyway.
Mike
anatech said:
So It will loose it's greatest quality, and will be... EXPENSIVE.
Hum, I don't think so.
What's the absolute maximum power supply voltage that I can apply?
I mean, what is the maximum supply with no signal in, and at full power that the amp can handle? Because if I apply 36v, at full power it will fall for something about 32v...
Hum, good (very good) to know this power / perceived volume relation. I'm surprised.
My Beyma speakers didn't arrived yet, so I'll start working in the original Symasym 5, and let's find out if I really need more power or not when the speakers arrive.
Thanks !!
Regards,
Fernando
Hi Fernando,
Many 100 W seakers can not actually handle that amount of power. I've seen my share of toasted voicecoils over time.
Hi Mike,
-Chris
That is the best idea.
Many 100 W seakers can not actually handle that amount of power. I've seen my share of toasted voicecoils over time.
Well, at least it will be more expensive for sure. That is unavoidable.
Hi Mike,
I fully agree with you. Low impedance loads using MJW devices may require another set. I'm liking these a lot.
-Chris
My symasym is quite happy driving my 4ohm speakers, but listening loud means output is seldom above 10v peak. (had scope attached while listening) This was enough to get frightening excursions in my vented speakers with deep bass parts. (Tschaikovsky digital canons, blue man group DTS-CD, what a fun ! With deep bass i mean really deep 
) My speakers should handle 100watts, but symasym would be able to destroy them mechanically. (with this music)
I already heard that klockkalockkrockklock, signaling that the magnet stopped the voicecoil from travelling further.
Mike
) My speakers should handle 100watts, but symasym would be able to destroy them mechanically. (with this music)I already heard that klockkalockkrockklock, signaling that the magnet stopped the voicecoil from travelling further.
Mike
-_nando-_ said:
Mike?
Well, there's no more reason not to start building the symasym
The MJE3281/1302 at farnell are quite cheap, and there's no expensive components at all, mpsa 18 I have lots of them, and I like it's sound, can be only my thought, but I can perceive that the sound of mpsa18 is better than BC549C.And, what is the final board? There's so many boards and eagle files that I'm lost, can you post the last board revision? (eagle files)?
Thanks !!
Kind Regards,
Fernando
Chris, my speakers are DIY (of course...) Not very high end. (Visatons, 2xW170S8 + G20SC, 17cm woofers, 2cm textile dome tweeter, 2 way, 40Liters, tuned at 36hz, and a real mans port 50cm^2x25cm, diy crossover,90db eff)
But, i am quite happy with them, uncolorized, brute deep bass, detailed.
20hz is too low for them... Max excursion 20mm.
Ooops, a bit dusty, time for some cleaning...
Fernando, i see max voltage at 40v. Get a good powerful toroid. maybe 250watts (at least), 2x25v.
You find all necessary data on my webpage, click on my www.
Mike
But, i am quite happy with them, uncolorized, brute deep bass, detailed.
20hz is too low for them... Max excursion 20mm.
Ooops, a bit dusty, time for some cleaning...
Fernando, i see max voltage at 40v. Get a good powerful toroid. maybe 250watts (at least), 2x25v.
You find all necessary data on my webpage, click on my www.
Mike
Attachments
Hi Mike,
They look good. 20 Hz is definitely too low. Box tuning B4 or other? You may consider making the port more resistive.
I still have my prototypes of a double box construction with 8" Marsland image 8's and a Philips AD0163T8. They sounded good, but much better with the Peerless 8" and Peerless tweeter. The box design was in Weems. Woofer control was very good. Bass extension is silly low (28 Hz -3dB with the Peerless 8"). The Marslands can't touch the Peerless. So they live in the bedroom.
They look good. 20 Hz is definitely too low. Box tuning B4 or other? You may consider making the port more resistive.
I still have my prototypes of a double box construction with 8" Marsland image 8's and a Philips AD0163T8. They sounded good, but much better with the Peerless 8" and Peerless tweeter. The box design was in Weems. Woofer control was very good. Bass extension is silly low (28 Hz -3dB with the Peerless 8"). The Marslands can't touch the Peerless. So they live in the bedroom.
My speaker building was before i joint here... so i down't know jack about "B4 tuning". Simple F-3db tuning.
Have you already tried the peerless xls10 ?
I also built a woofer with -3db@25hz. This is painful... I don't use it. My room/wall/ceil has a resonance at ~23hz, feels like an earthquake.
Everything standing/lying around starts to make strange noises.
About the port, i used it extra large to get least resistance.
Mike
Have you already tried the peerless xls10 ?
I also built a woofer with -3db@25hz. This is painful... I don't use it. My room/wall/ceil has a resonance at ~23hz, feels like an earthquake.
Everything standing/lying around starts to make strange noises.
About the port, i used it extra large to get least resistance.
Mike
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.