Inductances everywhere ...
Knock yourself out, keen .... read this , High Frequency Non-Inductance Melf (RFM) Resistors - Token Components Those resistors are good to several Ghz. A standard 100R mouser 1/4w resistor is good to 1ghz. Different sputtering techniques and oxide formulea have different characteristics and inductances. Bob cordell once told me on a triple OPS to add some inductances in series to the stoppers and between collector and rail to check real stability. For a good ol' bandwidth restricted audio amp, what we use for resistors is sufficient. I read wirewound's can have very high inductance , we also are driving the biggest inductor of them all (the woofer). At least my inductances are symmetrical. 😀
It IS good we are at least aware of this.
That also goes for using any other "exotics" (like wahibs sankens - I can't even find a source) , use what is cheap and available ... my sankens (2sc3263/sa1294) are used in the sherwoods , the ON's are $1.14 apiece , why bother with anything else ?
OS
Knock yourself out, keen .... read this , High Frequency Non-Inductance Melf (RFM) Resistors - Token Components Those resistors are good to several Ghz. A standard 100R mouser 1/4w resistor is good to 1ghz. Different sputtering techniques and oxide formulea have different characteristics and inductances. Bob cordell once told me on a triple OPS to add some inductances in series to the stoppers and between collector and rail to check real stability. For a good ol' bandwidth restricted audio amp, what we use for resistors is sufficient. I read wirewound's can have very high inductance , we also are driving the biggest inductor of them all (the woofer). At least my inductances are symmetrical. 😀
It IS good we are at least aware of this.
That also goes for using any other "exotics" (like wahibs sankens - I can't even find a source) , use what is cheap and available ... my sankens (2sc3263/sa1294) are used in the sherwoods , the ON's are $1.14 apiece , why bother with anything else ?
OS
Those are really cool looking resistors though...
In the direct signal/BW path they may not do much in terms of sonics, but It may be important in places with inherently high OLG loops. For instance, the two-Q CCS. Pretty simple, very short leads, but nonetheless I have seen these oscillating in simulation (w/parasitics). For this reason I sometimes include a 10R+10n zobel C-E of the current sense transistor.
MOSFETs obviously are another place. I read that a MOSFET's self-instability is proportional to gate inductance. So less inductance between gate and driver is better. But what gain comes from putting your driver 1cm from the MOSFET if your resistor already has an inch worth of parasitics?
Another thing to consider is immunity to RF. Even if the inductance doesn't affect circuit operation, it may increase susceptibility for areas of the circuit not BW-limited.
- keantoken
In the direct signal/BW path they may not do much in terms of sonics, but It may be important in places with inherently high OLG loops. For instance, the two-Q CCS. Pretty simple, very short leads, but nonetheless I have seen these oscillating in simulation (w/parasitics). For this reason I sometimes include a 10R+10n zobel C-E of the current sense transistor.
MOSFETs obviously are another place. I read that a MOSFET's self-instability is proportional to gate inductance. So less inductance between gate and driver is better. But what gain comes from putting your driver 1cm from the MOSFET if your resistor already has an inch worth of parasitics?
Another thing to consider is immunity to RF. Even if the inductance doesn't affect circuit operation, it may increase susceptibility for areas of the circuit not BW-limited.
- keantoken
Well in my book, IF you are not using 7 wrapped by 2 bound by 1 Silver coated MIL Spec Teflon 10 awg, then any amp will sound like muddy water...😀
And to think I was criticized for suggesting SMD for output emitter resistors not long ago. Go figure.😛
Happy New Year to ALL
And to think I was criticized for suggesting SMD for output emitter resistors not long ago. Go figure.😛
Happy New Year to ALL
1. - Yes , I have succeeded in getting a "2-Q" CCS to oscillate with my highest gain ksc1845's (U). I did this on purpose to see my "margin" in a real circuit. On my new AX/BX I either use 2- 2sc2229's of a "F" class 1845 pair "face to face" ... NO oscillation , A faulty IPS CCS is the one thing that would fry everything else after it .. no doubt.
2. - I am about to diverge from BJT's - in my PBM120 ( 2 pair IRF/lateral op's) , I intend on 270/330R full carbon gatestoppers in the 12mm between driver and MOSFET .. I ain't takin' any chances.
3. Yes , RF is the enemy. B-C shunt of 33-68pF on the drivers eliminates it's effect on a EF2 or triple BJT. In a grounded metal case with a cable RF filter , ferrites on the inputs-MOSFET gates ... the only RF you might see would be generated at 20mhz or above and by the mosfets themselves. 🙁
OS
Wow, just shocked myself. I hooked up an old carbon comp 5W 3Kohm resistor to my Fluke (5.5 digits) and watched as the resistance increased 2 whole ohms when I held it in my fingers. The same test with a 1/4W 100R carbon film reveals a negative tempco.
- keantoken
- keantoken
The carbon mixed with the ceramic varies wildly with temp. and humidity 🙁 . The Ni-cr film conducts just a little better with increasing temp. Just before they burn totally and turn glowin' red hot , a films resistance will drop drastically , then POOOFFFT...open circuit. 
OS
OS
Okay, noob here...
I'm looking for a DIY amp project, and this looks like an interesting one. And not outdated yet !
I haven't studied everything yet, but I'm assuming the AX board is better (more parts 😀). Any reason to go with BX ?
Can somebody tell me what "TMC" is ?
Any plans to make up a batch of boards ?
I'm looking for a DIY amp project, and this looks like an interesting one. And not outdated yet !
I haven't studied everything yet, but I'm assuming the AX board is better (more parts 😀). Any reason to go with BX ?
Can somebody tell me what "TMC" is ?
Any plans to make up a batch of boards ?
Oops, noob here again.
I figured out TMC : Transitional Miller Compensation.
It's been all of 18 hours since I read about it in Cordell's book.
Carry on.
I figured out TMC : Transitional Miller Compensation.
It's been all of 18 hours since I read about it in Cordell's book.
Carry on.
You make the boards (for now) I just spent my "lunch money" on new speaker parts , no FR-4 for me 🙁🙁 . The BX and AX both sound phenomenal , they just have a slight different "character". The best way to describe the difference is that the BX is a slightly better "easy listening amp" and the AX really shines at concert hall levels. Only with the 200 watt OPS of mine does one hear these very minor differences. On my PB60 , There is no perceivable difference.
I put both amps through hell on new years eve , even directly shorting the BX/PB60. Blew a fuse .. replaced it , it lived !
OS
Is there any reason that you have removed TMC from your latest BX files? I am still waiting for some transistors to turn up to finish mine, but am half way through assembling (Same version as SNG built).
I am also rebuilding the power boards (PB120) as I want to put the cap multiplier in now that I have 58v rails. Is the PB150 now finalized as I am about to put an order through to Farnell for parts.
Thanks pete
I am also rebuilding the power boards (PB120) as I want to put the cap multiplier in now that I have 58v rails. Is the PB150 now finalized as I am about to put an order through to Farnell for parts.
Thanks pete
BX really did not benefit as well from TMC as the current sourced version. Reason #2 is that I wanted my BX as the reference to carlo's "ES" to give me a sonic "baseline" to compare the AX with TMC to. The AX won , it is my living room amp. 🙂 The BX is still very nice to listen to , it will be my "outside" amp for parties and workshop. Both are extremely durable and fault tolerant , built to last 20+ years hard duty. 😎
PB150 ???? I have the 250H ....
Here is the PCB... (below) it uses the same schema as the 250N. This is my "baby" nicest BJT OPS I have ever "abused" 😀
OS
Sorry, I meant the PB120N. Theres no way I could need any more power than that. There seem to be quite a few changes to the power boards compared to the earlier versions. Just checking that everything is up to date before I place my order.
Would you still say it is worthwhile including the TMC in the BX, I already have the parts and boards made to include them?
Cheers Pete.
Would you still say it is worthwhile including the TMC in the BX, I already have the parts and boards made to include them?
Cheers Pete.
You will have to decide whether there is any audible advantages to TMC'ing the BX. I have mine (which was the last version board) with a 68pF c7 , C8 jumpered and the tmc resistor omitted. It still sounds wonderful. To explain why TMC "likes" the AX better , the BX already has OP feedback through the bootstrap and at the same time you have the tmc feedback. The AX , on the other hand ONLY has the TMC feedback loop as the VAS is biased by a fixed current source. The differences are minimal except at VERY high volumes , where TMC has more Xover distortion to actually correct. Feel free to experiment , that is why I specified in the schema for either TMC or no TMC. 🙂
I you want build an AX pair as well , then you can have 2 excellent amps in the house to "fuss" over. 😀 I am VERY happy with both and they are finalized (perfect).
Everything in my folders is now built and tested EXCEPT the PB120N . It is the exact same circuit as the 250N/H (I built 6 pairs of these!!) and should be flawless as well. Both power supplies (PS50/100 -built 7 of these) have been cycled 100's of times , and thoroughly abused. The PB60X2 and PB250H are my permanent house/barn amps and have also been cycled/abused. Absolutely guaranteed to work and amaze if built to spec.
OS
Last edited:
OS, now that they're perfect, I'm curious to see scope shots of the voltage between the input and output of the output stage (from the VAS side of the Vbe multiplier to the output at the emitter resistors) when playing at very loud levels. You may have one last hurdle to clear, thermals. Set your dual channel scope to differential mode if possible...
At louder volumes, output junctions heat up, decreasing Vbe, increasing bias, getting closer to class A (sounds good at high volumes...?) and thermal runaway (the Vbe multiplier will rarely compensate well, going under or over). (I have rudimentary thermal models for thermaltrak, output, driver and small-signal transistors if you're interested). Memory distortion exists in junction self-heating effects (memory distortion is a bad name), but simulators need "upgrades" to show this.
- keantoken
At louder volumes, output junctions heat up, decreasing Vbe, increasing bias, getting closer to class A (sounds good at high volumes...?) and thermal runaway (the Vbe multiplier will rarely compensate well, going under or over). (I have rudimentary thermal models for thermaltrak, output, driver and small-signal transistors if you're interested). Memory distortion exists in junction self-heating effects (memory distortion is a bad name), but simulators need "upgrades" to show this.
- keantoken
I will scope when i get new longer probes for my 20mhz , I am using a wire now for my scope work 🙁 How do you get a digital camera to take good scope shots ?? I must learn this.
You keep mentioning thermal stability to me ??? I have artificially went from outside in the cold to max hair dryer temp.I also have driven both amps in EVERY way , testing for mV along the way. In conclusion , I have never seen more than a +- 2mv deviation in either amp , on any of the emitter resistors , even after running for 5 hours strait with heavy metal
. I only have 2 pair of the Renesis 2sk1058/j162's and 2 pair to-3 laterals , these are for my sub amp , even as I will audition them with my full range setup.
I have built the symasym , quasi's amps , ampslabs projects , and several of ESP audio's diy projects , none have been as electrically and thermally stable as these. NONE have sounded like these. I just bought some nice Dayton soft dome tweeters and pyle 100w sealed back midranges. I will build a first order X-over w/ 600HZ and 4800Hz points to REALLY see what these amps sound like. I AM going to build some variant of the goldmund and artu's amp , possibly this month. I now realize , that the AX/BX will be nearly impossible to beat , soundwise , but I like to build ... just for the sake of building. 🙂
OS
Stability is obviously not a problem, but I'm talking about crossover distortion. At high volumes, in the simulator I see the bias increase for some time and then the multiplier begins to work and overcompensates, decreasing bias and increasing crossover distortion. In any case, the number of mV bias doesn't tell us that much. To actually SEE if there is any effect on linearity of crossover, you should scope it the way I suggested. This is also how I would set my bias currents for my amps, by visually inspecting the crossover region (tune for smallest bump at maximum output).
Just one more thing to test I guess, but your amps are already proven in the field...
- keantoken
Just one more thing to test I guess, but your amps are already proven in the field...
- keantoken
Am I looking for trouble and poor performance by trying to utilize run of the mill transistors for the Mongrel series of amplifiers ?
I have a bunch of the Slone semiconductors: 2N5551/ 2N5401/ 2SB649a/ 2SD669a/ MJ15003/ MJ15004. Board layout changes are a given.
I have a bunch of the Slone semiconductors: 2N5551/ 2N5401/ 2SB649a/ 2SD669a/ MJ15003/ MJ15004. Board layout changes are a given.