Member
Joined 2009
Paid Member
That's a cool feature, I'll have to take advantage of that.
Here's the BOM for the schematic of post 87.
By the way, yes I did use a huge amount of VAS degeneration. I borrowed this gem from Bryston. They use such a design approach in the integrated amp I own. It allows use of a higher collector load for the input devices which improves their gain - you don't lose quite as much gain as you might think because of the high VAS degeneration. But I'm not looking for really high OLG - some other people might like more OLG and hence higher negative feedback factors. I wanted to avoid that. The simulated FFT shows about equal H2 and H3 at relatively low levels and then the harmonics after that are down in the basement
Here's the BOM for the schematic of post 87.
By the way, yes I did use a huge amount of VAS degeneration. I borrowed this gem from Bryston. They use such a design approach in the integrated amp I own. It allows use of a higher collector load for the input devices which improves their gain - you don't lose quite as much gain as you might think because of the high VAS degeneration. But I'm not looking for really high OLG - some other people might like more OLG and hence higher negative feedback factors. I wanted to avoid that. The simulated FFT shows about equal H2 and H3 at relatively low levels and then the harmonics after that are down in the basement
Attachments
Last edited:
Member
Joined 2009
Paid Member
Thanks - I'll take advantage of this too !
Next up... I want to finish this project with a simple speaker dc-protection circuit. I've attached an LTSpice model of what I have in mind. I plan to use solid state relays and they are indicated in the spice model by three LEDs in series since this is how the SS relays work.
I'm planning to use SS relays because I have not seen mechanical relays with adequate ratings to be readily available. I will make my own SS relays.
I use both paralleled in my universal advanced protection circuit.
www.esperado.fr - Une protection pour amplificateurs universelle
(Googeule translate is your friend ;-)
Last edited:
Member
Joined 2009
Paid Member
Member
Joined 2009
Paid Member
Member
Joined 2009
Paid Member
Looks like aztec pottery
Nice amp BTW.
Yes it does !
Well I've populated the board - took about an hour. Photo attached. I am getting to like this surface mount stuff, it can be pretty quick and it uses hardly any solder (so less fumes and less flux to clean up afterwards). The dual-LED status indicator is mounted off-board on the front panel. I used one of those LEDs with '3 legs' that glows red or green depending on how you feed it. If it's green, all is good, if it's red then the dc protect has been activated.
I bench tested the unit - worked first time
I have installed it into the amplifier chasis and connected it up. As far as I can tell so far, it has no impact on the sound quality - good news.
I also tested it with a speaker attached and playing music. I used a somewhat expendable speaker and monitored the voltage across it with a DVM. I built the amp with dual RCA inputs per channel, one dc-coupled and one ac-coupled. I fed music into the ac-coupled RCA and then hooked up a bench supply via a resistor to apply dc to the dc-coupled RCA. This allowed me to apply varying levels of dc to the input of the amp and measure the output.
In the rather blurry attached photos you see one image of the set up where the amp has a green light, no dc applied. I applied dc and watched the cone move further outwards. With enough dc applied the protection was triggered and the output cut-off. The LED changed to red.
Attachments
Member
Joined 2009
Paid Member
Wow, I'm late again. I suppose I've been spending many months in forum hibernation, thinking about amplifier compensation... I'm working on my own design, not really revolutionary except that it uses a whole slew of tweaks. And often I come here, see something incredible and briefly consider scrapping my current prototype and starting over!!!
Bigun, have you seen my simulations of the SSA circuit? I think you should really try RC compensation. There is an easy way to compensate those CFP drivers if you have a signal generator.
http://www.diyaudio.com/forums/solid-state/193923-simple-symetrical-amplifier-49.html#post2988379
Later on I describe the "symasym" type compensation in detail. Look at the extreme reduction in high order harmonics in the FFTs later on, I suspect this is what you're looking for.
RC compensation, whether it's a snubber, stability comp or whatever, is a labor of DIY. It takes some trimming, but it's worth it.
Also, get that line filter. Snub your rectifier properly and also try snubbing your trafo secondary. For instance:
http://www.diyaudio.com/forums/aksa/191053-swordfishy-aspen-fetzilla-power-amp-19.html#post2998407
In my prototype it made an incredible difference. Try it at least once.
Bigun, have you seen my simulations of the SSA circuit? I think you should really try RC compensation. There is an easy way to compensate those CFP drivers if you have a signal generator.
http://www.diyaudio.com/forums/solid-state/193923-simple-symetrical-amplifier-49.html#post2988379
Later on I describe the "symasym" type compensation in detail. Look at the extreme reduction in high order harmonics in the FFTs later on, I suspect this is what you're looking for.
RC compensation, whether it's a snubber, stability comp or whatever, is a labor of DIY. It takes some trimming, but it's worth it.
Also, get that line filter. Snub your rectifier properly and also try snubbing your trafo secondary. For instance:
http://www.diyaudio.com/forums/aksa/191053-swordfishy-aspen-fetzilla-power-amp-19.html#post2998407
In my prototype it made an incredible difference. Try it at least once.
Member
Joined 2009
Paid Member
Those are some very sound ideas, certainly worthy. Trouble is, I'm knee deep in other projects.
One day I may come back to TGM5 and do some tweaking. The kids were in the basement playing an iPod through the TGM5 and quite enjoying it !
I do need to revisit the dc-protection as it gives me a slight turn-off thump where none existed before and that's the only thing I changed although I haven't worked out what the cause is.
I usually come back to revisit my projects after awhile. My TGM1 was the first amp I built and it got completely recycled this year into a new box with new PSU and sounds good.
But right now I'm playing with tubes again - I keep coming back to tubes, there's something about the sound of tubes that I have not been able to hear with SS yet.
One day I may come back to TGM5 and do some tweaking. The kids were in the basement playing an iPod through the TGM5 and quite enjoying it !
I do need to revisit the dc-protection as it gives me a slight turn-off thump where none existed before and that's the only thing I changed although I haven't worked out what the cause is.
I usually come back to revisit my projects after awhile. My TGM1 was the first amp I built and it got completely recycled this year into a new box with new PSU and sounds good.
But right now I'm playing with tubes again - I keep coming back to tubes, there's something about the sound of tubes that I have not been able to hear with SS yet.
I'd like to offer an opinion on that
Texture, like "needle on a record" and tape as well
and
Harmonic texture, like not really a soft clipper, and Moosefet IRF510 preamp for triode like harmonics.
and
Lightspeed, much like resistors in series for base stoppers, blarebuster, capacitive headroom booster and other impedance hijinks.
and
Decreased feedback current (it doesn't always need max) and Dial-A-Tube output transformer simulator.
and
Capacitor signature texture--Elna Cerafine 4.7uF//tiny polyster dip cap combination as an input cap is today's available recipe for a classic part.
Texture (esp triode type) breaks/cracks droning tones to facilitate easier hearing, and some impedance modifications slack/decrease compression to facilitate easier hearing. It is also done to make a "sounds good loud" system actually sound good loud (non-blaring). It can also put the finishing touches on MP4/aac and some other digital sources, for the purpose of enhancing intelligibility.
The fun thing about doing this kind of work in solid state is the probability to have extraordinarily high resolution presented together with classic textures/harmonics, as they are not mutually exclusive. Solid state also gives you the chance to make your effects switchable and variable so you can dial in the needed amount, and a needed amount is indeed only as much as necessary to cancel some other problem, namely you want to cancel blare so that you can hear "through" and indeed that works ideally with low noise solid state.
You're lucky if you don't have problems with tubes. I think you have worked more with bipolars. Bipolars are too clinical, too far from tube characteristics. Mosfet is closer. Class-A unfortunately, because class-B tend to bring fatigue that is not exist with (all simple) tube amps. If you try to bias hexfet higher until you cannot hear sound improvement, you will find out that the bias current is way above 4A
Hey Daniel. I'm interested in what you think of the texture of these recordings:
Info Nympho | Cuddle Magic
More info:
Cuddle Magic To Release "Info Nympho", An Analog Record by Cuddle Magic — Kickstarter
It seems not unlike that of these:
http://www.soundkeeperrecordings.com/americas.htm#samples
http://www.soundkeeperrecordings.com/confluence.htm#samples
Except the Info Nympho recordings seem clearer, less colored. In all of these recordings the signers and instruments sound as if they harmonize perfectly, I don't understand whether this is the skill of the singers or an effect of the recording equipment. I like it though.
Info Nympho | Cuddle Magic
More info:
Cuddle Magic To Release "Info Nympho", An Analog Record by Cuddle Magic — Kickstarter
It seems not unlike that of these:
http://www.soundkeeperrecordings.com/americas.htm#samples
http://www.soundkeeperrecordings.com/confluence.htm#samples
Except the Info Nympho recordings seem clearer, less colored. In all of these recordings the signers and instruments sound as if they harmonize perfectly, I don't understand whether this is the skill of the singers or an effect of the recording equipment. I like it though.
Last edited:
Cuddle magic, like a tube radio with a tone control dial. Success depends on both replay equipment and taste in music. I suspect that it will be more successful to replay it on LP.
Confluence, their digital fuzz is a little brighter than the real thing and nice for a brief listen but not great for a long listening session. It seems to be recorded somewhat like pop music for broadcast. These tracks could be cleaned of that noise almost losslessly since it is a predictable digital effects pattern.
I think that some of that shows what you don't want your power amp to do. Effects on a recording changes with each recording. Effects on a power amp may be like a nonstop poke in the ear, because there's a listening fatigue possibility if effects are built into a power amp, since that would not change per each track.
Americas, just like one mosfet. I like them except that they fuzzed the sax a little bit for a more air less horn effect. Otherwise, very nice and clear. Not overstated. It is a good recording. This would probably not cause listening fatigue.
P.S.
It would be important to mention that Bigun isn't going for texture--he wants soundfield depth. Mosfet is suitable for both though--both at once if you like. Of course it is doable with BJT, but maybe just easier with mosfet.
Last edited:
Hey Daniel.
I agree with what you say about what is wanted from a power amp. Right now my design philosophy is that the amp should play the music close to the way the recording artist heard it, unless this paradoxically involves increasing distortion. I think niceness effects should be added during recording by the artist, or by a defeatable line effect.
Soundfield depth is also what I go for. I believe that emotional engagement is important and if the amp takes that away, that qualifies as a distortion (although there IS music I'd rather not be emotionally engaged to). If distortion can't be eliminated however, we have to choose which kind of distortion we can live with. I am concerned mostly with 3 things: fatigue, realism in general, and emotional engagement.
I agree with what you say about what is wanted from a power amp. Right now my design philosophy is that the amp should play the music close to the way the recording artist heard it, unless this paradoxically involves increasing distortion. I think niceness effects should be added during recording by the artist, or by a defeatable line effect.
Soundfield depth is also what I go for. I believe that emotional engagement is important and if the amp takes that away, that qualifies as a distortion (although there IS music I'd rather not be emotionally engaged to
). If distortion can't be eliminated however, we have to choose which kind of distortion we can live with. I am concerned mostly with 3 things: fatigue, realism in general, and emotional engagement.I would change that to the way the recording artist and his producer wants you to hear it.
That could be identical or a little different or a lot different from what is actually heard "live" in the studio.
It is for this reason I never use tone controls.
I want to get as close to the way the artist wants me to hear it, not some distorted version that I can concoct.
Last edited:
I don't know how to relate them directly to electronics. But, I did come up with three circuit areas to explore.Soundfield depth is also what I go for. I believe that emotional engagement is important and if the amp takes that away, that qualifies as a distortion (although there IS music I'd rather not be emotionally engaged to
Power circuit clues?
I wonder and I guess, that applying a filtered transformer, some soft schottky just for sport at the power supply, and also consider capmulti/ripple eater at small signal could help produce the desired effects in TGM5. If the main feedback has to work to remove noise, it may concurrently reduce any audio that is alike the noise. So, maybe, probably, uber-clean power first, just to find out if and how much that helps?
Small signal clues?
If that won't do it, then perhaps exploring more or less feedback current may help reveal something about the amplifier. With this new topology, there may be compensations located unusually, as if hidden from view, but they're not hidden from the feedback and thus the optimal current within the feedback resistor and also the gain setting as well, need additional exploration. The SSA thread recently replaced some passive components with CCS modules for the purpose of more open sound. Since that worked, there's definitely something going on at small signal current. I'm just reporting that a clue exists.
Large signal clues?
I believe that Nico's home amplifier is almost the same as TGM5 except for the output section is fet. Since several SSA Fet output editions are high achievers, there may be a difference that needs to be found and that difference replicated with BJT for the "all BJT" TGM5. What is this audio difference and how do we install it without installing a Fet?
What do you think?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.