Member
Joined 2009
Paid Member
Member
Joined 2009
Paid Member
you're enjoying yourself too much !
With Carlos going commercial now I can't just 'give it away' 😀
With Carlos going commercial now I can't just 'give it away' 😀
No business for a while dear Bigun... i will continue DIY, for free, and those
amplifiers, offered for free in this forum, as DIY, non profit amplifiers, will continue to be this way from now to the ethernity...or till the day music will be transmitted in multi channels by Geo stationary Satelites and reception will be an implanted chip under your skin at the top or your head and connected using golder wires directly to your brain's listening area.
To go to Pro is a plan to July... and the amplifiers offered will be different, to other group of people...with WEB advertising and maybe i will have banner here too..depends on cash flow....not a pro till 2011 and July.
Dx amplifier (tradicional generic bootstrap amplifier from the seventies)
HRII - natural evolution to the Dx amplifier
Precision 1 - a higher power HRII having some interesting circuitry
DHR Turbo - a Dx amplifier with steróids
Dx Blame models - a Blameless with some personnal touches
Troyan - an amplifier for home demolition purposes.
And some others too.... the ones i had gerbers, they were offered to the community and published..so..they are public dominium amplifiers.
All these amplifiers will be connected to DIY, for ever an ever and to this nice forum.
I will be offering help and follow up to the ones have bougth Blue boards to Supercharged and other models, offered in the group buy..they were already shipped world wide.
Alike Hugh Dean and others... i will be at the forum as a reader and mine future amplifiers will not be mentioned or advertised.... so... i will continue to be uncle charlie while surfing the forum and will have another activity but outside the forum...the same you see around with lot's of people.
My complain, in the past, was to offer almost the same...or something not too much different...better to have hardly modified things..about to be copied or "inspired"..no problems to me...as mine are also copied and inspired..it is impossible in modern times not to be influenced..... with high speed internet connections and all technology at the tip of our fingers...we need just to type and the informations arrive...we cannot be too much original..i am not and i do think nobody is original anymore..we are the result of influences we receive from others.
We had a dark age in our forum..along some monthes all amplifiers where bootstrapped...was terrible...only one style going on.... i am glad this finished and we have many other styles.
Do not think on me as different..i am still uncle charlie..with the same old heart..and that story of Pro depends on a lot of things to come true.
An skilled professional told me:
- "They will never forgive you if you go pro"
I hope this will not be truth... as people usually dislike when you become "different" from the group.... this is not very well accepted.
regards,
Carlos
amplifiers, offered for free in this forum, as DIY, non profit amplifiers, will continue to be this way from now to the ethernity...or till the day music will be transmitted in multi channels by Geo stationary Satelites and reception will be an implanted chip under your skin at the top or your head and connected using golder wires directly to your brain's listening area.
To go to Pro is a plan to July... and the amplifiers offered will be different, to other group of people...with WEB advertising and maybe i will have banner here too..depends on cash flow....not a pro till 2011 and July.
Dx amplifier (tradicional generic bootstrap amplifier from the seventies)
HRII - natural evolution to the Dx amplifier
Precision 1 - a higher power HRII having some interesting circuitry
DHR Turbo - a Dx amplifier with steróids
Dx Blame models - a Blameless with some personnal touches
Troyan - an amplifier for home demolition purposes.
And some others too.... the ones i had gerbers, they were offered to the community and published..so..they are public dominium amplifiers.
All these amplifiers will be connected to DIY, for ever an ever and to this nice forum.
I will be offering help and follow up to the ones have bougth Blue boards to Supercharged and other models, offered in the group buy..they were already shipped world wide.
Alike Hugh Dean and others... i will be at the forum as a reader and mine future amplifiers will not be mentioned or advertised.... so... i will continue to be uncle charlie while surfing the forum and will have another activity but outside the forum...the same you see around with lot's of people.
My complain, in the past, was to offer almost the same...or something not too much different...better to have hardly modified things..about to be copied or "inspired"..no problems to me...as mine are also copied and inspired..it is impossible in modern times not to be influenced..... with high speed internet connections and all technology at the tip of our fingers...we need just to type and the informations arrive...we cannot be too much original..i am not and i do think nobody is original anymore..we are the result of influences we receive from others.
We had a dark age in our forum..along some monthes all amplifiers where bootstrapped...was terrible...only one style going on.... i am glad this finished and we have many other styles.
Do not think on me as different..i am still uncle charlie..with the same old heart..and that story of Pro depends on a lot of things to come true.
An skilled professional told me:
- "They will never forgive you if you go pro"
I hope this will not be truth... as people usually dislike when you become "different" from the group.... this is not very well accepted.
regards,
Carlos
Last edited:
Member
Joined 2009
Paid Member
Well now, I have completed the modifications. The output device emitter resistors paralleled up to achieve Re of 0.2R, I increased the current flow slightly through the LTP master devices to bring it up from 0.8mA to over 1mA, I removed the 'TGM goes tubey' magic resistors, I drastically reduced the phase lead compensation (from 22pF to a few pF) and increased the LTP degeneration to 100R. This last move cuts down the OLG a bit, but the lower value Re helps reduce distortion at the output to compensate for the reduction in negative feedback. I think it's a good tradeoff. Had to reset the biass to around 20mA across the emitter resistors (which may be a tad low as 80mA would require 32mV).
I also tried installing the VAS pre-buffer and managed to do it neatly. I found that it demands much higher phase lag compensation and although I beefed it up I still had poor bias stability. I've seen this before - stability is on the hairy edge if the bias wanders around, I don't even have to fire up the scope. So I've backed off using a VAS pre-buffer on this amp, it will have to be revisited another time.
I've set it up as a stereo unit fed with my YBA CD player. I found myself sat there listening to it in my 'lab' and had forgotten I was supposed to be doing something constructive. I realized that this amp sounds darn good and it had carried me away with the music. It's a sign that I need to stop meddling with it.
The reason I'm able to have TGM2 up on the bench again is because it's been replaced with a commercial home theatre receiver that I received as a gift a couple of weeks back - but I'm not sure it's going to be up to the task after listening to TGM2 - I think I'm going to have to start building from new again....
I also tried installing the VAS pre-buffer and managed to do it neatly. I found that it demands much higher phase lag compensation and although I beefed it up I still had poor bias stability. I've seen this before - stability is on the hairy edge if the bias wanders around, I don't even have to fire up the scope. So I've backed off using a VAS pre-buffer on this amp, it will have to be revisited another time.
I've set it up as a stereo unit fed with my YBA CD player. I found myself sat there listening to it in my 'lab' and had forgotten I was supposed to be doing something constructive. I realized that this amp sounds darn good and it had carried me away with the music. It's a sign that I need to stop meddling with it.
The reason I'm able to have TGM2 up on the bench again is because it's been replaced with a commercial home theatre receiver that I received as a gift a couple of weeks back - but I'm not sure it's going to be up to the task after listening to TGM2 - I think I'm going to have to start building from new again....
Hi Bigun,
pretty nice implementation (in post #132). I would remove Q42 doing nothing good, likewise C28, increasing the value of C31, would insert emitter resistor for Q29 say 100 Ohms, C24 should be directly parallel with R48.
Merry Christmas.
pretty nice implementation (in post #132). I would remove Q42 doing nothing good, likewise C28, increasing the value of C31, would insert emitter resistor for Q29 say 100 Ohms, C24 should be directly parallel with R48.
Merry Christmas.
Bigun, if you're up to it, you could also try moving the top of C27 to the emitter of Q25. This will remove some odd harmonics.
- keantoken
- keantoken
Member
Joined 2009
Paid Member
Hi WuYit, Yup, Q42 has been removed but although we all love to hate Cdom (C28) it's an item of necessity in the Lin Topology and in practice I have found that sonics are worse with a value that's too low. Perhaps there are better topologies to explore in this regard (a hint as to what I'm working on in the background). It's a nice idea to add some degeneration to the VAS device, although 100 Ohms is quite a bit. What I found interesting was your comment on C24 - it's not clear to me why this capacitor should parallel R48 exactly. If we look at the purpose of C24 it's for 'charge suck out' where direct connection to the base of the output devices makes sense.
Hi kean - I've been finding it difficult to get a good correlation between spice fft and sound, at least it seems ok for gross first order effects but after that it's increasingly difficult. Making this change would be difficult on the pcb as the capacitors legs are underneath the body of the cap and soldered directly to the relevant traces, the surgery might not be very pretty so before this could be contemplated can you tell me - did you try this change before and get an audible result ?
Carlos - I will look forward to seeing your continued presence, this forum would not be the same without your enthusiasm ! but if you are only cruising this place and not making amplifiers here, well, it just won't be the same. It won't do at all !
Hi kean - I've been finding it difficult to get a good correlation between spice fft and sound, at least it seems ok for gross first order effects but after that it's increasingly difficult. Making this change would be difficult on the pcb as the capacitors legs are underneath the body of the cap and soldered directly to the relevant traces, the surgery might not be very pretty so before this could be contemplated can you tell me - did you try this change before and get an audible result ?
Carlos - I will look forward to seeing your continued presence, this forum would not be the same without your enthusiasm ! but if you are only cruising this place and not making amplifiers here, well, it just won't be the same. It won't do at all !
Last edited:
I haven't tried it - never had the chance. The theory is pretty sound though. Taking the outputs out of the bootstrap loop makes the bootstrap MUCH more effective at its basic task, and it only injects harmonics produced by Q25. What surprises me is how many people refused to even try it when I asked. When I FINALLY get around to making an amp of my own it'll be in the design default.
One reason LTSpice result may not correlate well with the sound is that LTSpice doesn't simulate thermal Vbe effects. I am working on this as of yet:
http://www.diyaudio.com/forums/solid-state/169590-mongrel-supersym-ii-22.html#post2407669
I have a thermal simulation of your amp running. If you want it, send me an Email or a PM so I can send it to you (I am keeping it home until I think it's ready). In one amp I simulated, thermals made the difference between dominant 2nd or 3rd harmonics.
Sometimes the output transistors can oscillate by themselves. This is because inductive parasitics can cause negative impedance in one pin of the device and if this pin sees a capacitor, it becomes an oscillator (negative impedance+capacitor=oscillator). Thus we use base resistors to cancel out this negative impedance. So if you want to be safe, you should at least put a resistor in series with the suckout capacitor.
BTW, my simulation of your amp included parasitics, and for some reason it likes to oscillate at a few hundred MHz unless I use a faster device like the 2SC3423 as the VAS.
- keantoken
One reason LTSpice result may not correlate well with the sound is that LTSpice doesn't simulate thermal Vbe effects. I am working on this as of yet:
http://www.diyaudio.com/forums/solid-state/169590-mongrel-supersym-ii-22.html#post2407669
I have a thermal simulation of your amp running. If you want it, send me an Email or a PM so I can send it to you (I am keeping it home until I think it's ready). In one amp I simulated, thermals made the difference between dominant 2nd or 3rd harmonics.
Sometimes the output transistors can oscillate by themselves. This is because inductive parasitics can cause negative impedance in one pin of the device and if this pin sees a capacitor, it becomes an oscillator (negative impedance+capacitor=oscillator). Thus we use base resistors to cancel out this negative impedance. So if you want to be safe, you should at least put a resistor in series with the suckout capacitor.
BTW, my simulation of your amp included parasitics, and for some reason it likes to oscillate at a few hundred MHz unless I use a faster device like the 2SC3423 as the VAS.
- keantoken
When you attempt to use a too low value for cdom, the amp will probably be unstable. The fact that the 'sonics are worse' is due to high frequency instability that is very undesirable and yes, possibly audible as 'brightness'.
Using a FG and scope (and giving your ears a rest, they are NOT high precision measurement apparatus) you can determine the ideal value for cdom.
If Cdom is too low, the amp will oscillate at certain output voltages, though it might be stable at 0V. This can sound like clipping.
- keantoken
- keantoken
Member
Joined 2009
Paid Member
Hi Kean, well I shall have to make a note of the bootstrap mod. I'm not sure what the fate of the boards will be yet - they were doing duty in my home theatre set up but I've replaced them with a commercial unit (the jury is still out on this, my wife and I feel sound quality has been traded off for convenience). I also have a need for a high quality desktop amplifier and I may build these boards into a new stereo option for that purpose. I have also offered them to a friend. So I don't want to bugker them up with more experiments until I know where they are going.
I've PM'd you on the simulations - I'd be interested to see how you've tackled the parasitics. I test all my amps with CRO to ensure no oscillations under normal use. I'm also very fortunate that the TGM2 design draws heavily on the knowledge of Hugh (I'm using a higher value of Cdom than Hugh recommends - just to be on the safe side) so I have absolute confidence in it, compared with my TGM3, which is mostly home-grown.
Cdom seems to affect sonics even without producing oscillations. I found this out sometime ago, I think it was on TGM1 but I can't remember (it's in the thread somewhere). Not only were lower values a bit 'bright' but higher values were a bit 'flat' and there did seem to be a range which produced the best sound. Unfortunately John, my body stopped being a high precision instrument some years ago...
I've PM'd you on the simulations - I'd be interested to see how you've tackled the parasitics. I test all my amps with CRO to ensure no oscillations under normal use. I'm also very fortunate that the TGM2 design draws heavily on the knowledge of Hugh (I'm using a higher value of Cdom than Hugh recommends - just to be on the safe side) so I have absolute confidence in it, compared with my TGM3, which is mostly home-grown.
Cdom seems to affect sonics even without producing oscillations. I found this out sometime ago, I think it was on TGM1 but I can't remember (it's in the thread somewhere). Not only were lower values a bit 'bright' but higher values were a bit 'flat' and there did seem to be a range which produced the best sound. Unfortunately John, my body stopped being a high precision instrument some years ago...
Last edited:
My observations are the same with any stability capacitor. I use a different type of compensation for my headphone amp and it's the same way.
Obeying the 25nH/inch rule, MJL3281/2SC5200 have about 20nH lead inductance at full lead length. You will be surprised how much difference this makes. TO92 and TO126 - 15nH.
I am having to write some small documentation, so it may be 30 minutes or more.
- keantoken
Obeying the 25nH/inch rule, MJL3281/2SC5200 have about 20nH lead inductance at full lead length. You will be surprised how much difference this makes. TO92 and TO126 - 15nH.
I am having to write some small documentation, so it may be 30 minutes or more.
- keantoken
Yes, overcompensation may be audible as well. Aiming for the correct value will put it in the range you speak of, where it sounds best.
Member
Joined 2009
Paid Member
Hi Kean - I'm expecting a very complex model which may be above my head anyway so don't put too much faith in me on this - I'm thinking it's not only lead inductance but you have the non-idealized behaviour of real world components to include - especially the load. Then of course you have the pcb layout to factor in.
The model is actually rather simple, and it just wraps around the existing transistor model.
I haven't gone so far as to simulate all parasitics. But just adding uncoupled lead inductances gets you much closer. On that note, I did make my own speaker model of the Wild Burro BetsyK... I suppose I will have to include it now. 😉
- keantoken
I haven't gone so far as to simulate all parasitics. But just adding uncoupled lead inductances gets you much closer. On that note, I did make my own speaker model of the Wild Burro BetsyK... I suppose I will have to include it now. 😉
- keantoken
Not much, in my experience. Parasitic capacitance and inductance in the typical audio amplifier layout is negligible and will most likely have no bearing on stability.
I suppose it depends on the type of amplifier you intend to create. My headphone amp has bandwidth out to 100KHz, and for this it needs a gain-bandwidth product of >10MHz. An open-loop gain of around 80db, tends to make small parasitics and resonances rather large.
- keantoken
- keantoken
- Status
- Not open for further replies.