Shawna, thanks so much for the offer and taking the time to go through the posts in this thread. You may kindly mail me a copy of the book at either or both of these ids: cred@rediffmail.com ///// cred_audio@rediffmail.com
Russ, thank you for the offer to keep looking. I look forward to hearing from you as well.
To the rest, I have built one channel of Randy's OPTIMOS circuit that used to be on the web some years ago. It is NOT the totem-pole design which he sent me later while still under development. It worked perfectly. I still have the board. I will get another board done and assembled to check if the bias problem, as reported by others does occur. By the way, the output stage is K1058 and J162. More on this as I made progress in some weeks.
The ready-made PC boards with the ZUS logo are tested to work perfectly. Both the OPTI-MOS as well as the TOTEM-POLE designs. If the amp does not appear to be working, it will most likely be a component misplacement and/or physical wiring errors. These amps are truly fantastic!! Enjoy the music.
Russ
Thanks Russ. I mentioned that the one channel of the OPTIMOS (not Totem Pole) which was on the Seal Electronics site, worked with a PCB design that I laid out. I want to try one more channel in order to complete a stereo amp. Someone here mentioned that since the VAS is 'floating', the amp would bias one day correctly and another day it would not.
However, in terms of sonics how do the original OPTIMOS and the TOTEM Pole compare?
However, in terms of sonics how do the original OPTIMOS and the TOTEM Pole compare?
Hello,
If I am not mistaken. Neither the TOTEM pole, neither the OPTIMOS are concerned by the VAS Iq problem. As for the optimos he does not use current mirrors. For the totem pole I guess neither... the only problem is when he uses Current Mirrors in conjunction with a mirror image topology. And certainly if he also uses a darlington in the VAS.
My experience is that with that combination component tolerances are crazy. 1% change of a resistor causes a 133% impact on the VAS Iq. Going 1% up the VAS Iq raises by a factor 10 (smoke is sure!) Going 1% down -> VAS Iq reduced by 10x (class C operation).
You should by 1% components or even better 0,1% ANd with a very low temp coeff. Measure before placing each resistor (i talk about the 8 degen resistors of the diff Qs and the CM Qs) in order to have the desired VAS IQ -> even here expect up to 30% change in VAS IQ due to resistor temp coeff. Since the CM degen res and diff amp degen res have inverse effect on the vas iq a global warming would eliminate itself ... however they might not compensate the same since one has more impact then the other... AND if the heat is not equal (eg you placed some resistors near a heated Q and another not) the result will be bad.
All this speech is without talking about the Q's themselves. Also here they have to be exactly matched and you will suffer thermic variations with great effect upon VAS Iq. Also here one is compensation opposingly to the other so literally tying them back to back together might reduce problems...
I think It is possible but very hard and dangerous for real speakers
Surely it will never be commercialized as manufacturers cannot accept to fiddle trial and error components or measure them one by one and see the behavior. Tying Q together is also problematic for them ... it is too ARTISANAL ... it cannot be automized. And then there is the guarantee ... they cannot take the risk. You for yourself could ... but i did it for years and I smoked A LOT OF Qs ... I abandonned this circuit ... but again its not the totempole neither the optimos circuit ...
Greetz
Sorry for the crazy wrting style but i am in a hurry
If I am not mistaken. Neither the TOTEM pole, neither the OPTIMOS are concerned by the VAS Iq problem. As for the optimos he does not use current mirrors. For the totem pole I guess neither... the only problem is when he uses Current Mirrors in conjunction with a mirror image topology. And certainly if he also uses a darlington in the VAS.
My experience is that with that combination component tolerances are crazy. 1% change of a resistor causes a 133% impact on the VAS Iq. Going 1% up the VAS Iq raises by a factor 10 (smoke is sure!) Going 1% down -> VAS Iq reduced by 10x (class C operation).
You should by 1% components or even better 0,1% ANd with a very low temp coeff. Measure before placing each resistor (i talk about the 8 degen resistors of the diff Qs and the CM Qs) in order to have the desired VAS IQ -> even here expect up to 30% change in VAS IQ due to resistor temp coeff. Since the CM degen res and diff amp degen res have inverse effect on the vas iq a global warming would eliminate itself ... however they might not compensate the same since one has more impact then the other... AND if the heat is not equal (eg you placed some resistors near a heated Q and another not) the result will be bad.
All this speech is without talking about the Q's themselves. Also here they have to be exactly matched and you will suffer thermic variations with great effect upon VAS Iq. Also here one is compensation opposingly to the other so literally tying them back to back together might reduce problems...
I think It is possible but very hard and dangerous for real speakers
Surely it will never be commercialized as manufacturers cannot accept to fiddle trial and error components or measure them one by one and see the behavior. Tying Q together is also problematic for them ... it is too ARTISANAL ... it cannot be automized. And then there is the guarantee ... they cannot take the risk. You for yourself could ... but i did it for years and I smoked A LOT OF Qs ... I abandonned this circuit ... but again its not the totempole neither the optimos circuit ...
Greetz
Sorry for the crazy wrting style but i am in a hurry
The design that doesnt work is the one shown in his book where the LTP s are current mirror loaded. In that design the vas bias is undefined and you can be sure it doesnt work correctly, some members here have developed workarounds to that circuit. The vas bias is dependent on device properties and what you get is very unstable vas bias changing according to temperature and level of input.
The optimos is a different design altogether where the current mirrors have been substituted with cascodes. Theres no problem with this.
The optimos is a different design altogether where the current mirrors have been substituted with cascodes. Theres no problem with this.
Sonically, they both are almost impossible to tell apart - if you even could. The TOTEM-POLE specs out better, but we are talking about mill-percentages of THD. I really do not believe one can hear these diferences. Perhaps those among us who can hear differences between resistors...
You see where I am going with tihs.
However, to me my choice for the ultimate expression is the TOTEM-POLE. It just "feels right". The best-of-the-best.
I really appreciate your inputs and candid views. The OPTIMOS circuit that I have is indeed cascoded symmetrical inputs, two-pole compensated cascode VAS and two pairs of lateral mosfet outputs, powered with +-65volt rails and input protected by back-to-back connected 3v9 zeners. The schematic says revision 2.0 dated 19 Oct 01. During our online discussions, Randy said that he had modified this very design to work as an ultrasonic power driver for some medical purposes (to drive 600KHz sine input).
He also sent me the TOTEM Pole design during development and on completion. I have the PDF of the schematic and board overlay, but do not have either the track pattern for the double sided board he used, nor the gerber files. I think they were property of ZEUS. So I may have to roll out my own single sided board, to hear what Russ has whet my appetite to.
He also sent me the TOTEM Pole design during development and on completion. I have the PDF of the schematic and board overlay, but do not have either the track pattern for the double sided board he used, nor the gerber files. I think they were property of ZEUS. So I may have to roll out my own single sided board, to hear what Russ has whet my appetite to.
Randy was a good person. A while back I took a plan out of his earlier book on basic electronics; was experimenting and changing things around, kept cornering myself and getting stuck. Called him and he was really helpful, a good listener, came across as a calm, peaceful sort of person. Applying his suggestions got things moving again.
I'd be interested in reading Randy's working notes on designing and building amplifiers, field testing them, lessons learned. He was an intelligent man and probably kept good documentation. (Actually that happens here every day in real time with contributors' projects)
I also would also be very interested in seeing any such technical career autobiography from at least three or four career designers here on diyaudio. Notes and papers organized in one place: for reference in building and troubleshooting, for letting the reader benefit from the other's learning curve and understand the effort, and for letting the reader get a close look at some higher-stakes problem solving in action.
Gads! Did I just forget that Mr. Pass has effectively done this for years by sharing his project papers? Thank you!
I'd be interested in reading Randy's working notes on designing and building amplifiers, field testing them, lessons learned. He was an intelligent man and probably kept good documentation. (Actually that happens here every day in real time with contributors' projects)
I also would also be very interested in seeing any such technical career autobiography from at least three or four career designers here on diyaudio. Notes and papers organized in one place: for reference in building and troubleshooting, for letting the reader benefit from the other's learning curve and understand the effort, and for letting the reader get a close look at some higher-stakes problem solving in action.
Gads! Did I just forget that Mr. Pass has effectively done this for years by sharing his project papers? Thank you!
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.