anatech said:
If you can bear with me Chris and not take it personally, but it is this kind of empiricism that leads to folklore. I would expect you would get the same result if you matched Vbe, in fact that is what you found. You also dramatically changed the thermal coupling. Unless you eliminate the cases of matched beta and unmatched Vbe (you would probably have to get very different date codes for that) and the case of matched Vbe's and unmatched beta you can't really make a definitive conclusion. Or maybe the topology is sensivive to the beta/Early-voltage product? This would tend to be constant for a given device.
BTW I bought a 100 4401/4403's on tape and the beta's and Vbe's sequentially off the tape are amazingly close.
janneman said:
Sorry Jan, I can't? Why?
What I can't do is to reproduce the POWER of that organ in a space the size of the one that the organ sits in, nor can I reproduce the same localization of sound that a real organ has via two channels of home audio.
What do you think the max SPL of an organ is in a typical listening position??
_-_-bear
Hi Scott,
Hey, no problem. As I stated, most of what I have learned comes from what I see in the field related to what I learned at Ryerson (in Toronto). I am certainly open to adjust my viewpoints on what I believe as better data or understanding comes my way.
I have very recently been trying to come to grips with whether the beta or VBE is what I should be looking at. I do suspect that it may be a combination, but since beta varies more than VBE, it may be the thing we should watch. I don't know.
In fact, I am shortly going to be doing some tests on this very subject. I sure hope I'm right! Either way, I'll learn something and I feel that is worthwhile in of itself. What I can tell you is that I have been matching transistors for many years. I know it makes a huge difference, but by matching beta, I may be automatically matching VBE. I also know that buying in 100 lots greatly enhances my yield of matched devices (by beta). I will bet that matching on VBE would increase my number of matched pairs.
I have attached the jig I use for matching these parts. The method removes the device temperature as a variable as two parts are compared, in contact with each other. A foam cover goes over the two transistors. It works so well that any mismatch in degeneration resistors in the target location makes a big difference. That suggests that I'm on the right track (again, I sure hope so!). I may be barking up the wrong tree here, so correcting my view will cut down on how silly I feel. 🙂
The matcher for PNP or P channel is the same, inverted polarities. Very quickly built, I should use a PCB as I do use it a lot.
-Chris
Hey, no problem. As I stated, most of what I have learned comes from what I see in the field related to what I learned at Ryerson (in Toronto). I am certainly open to adjust my viewpoints on what I believe as better data or understanding comes my way.
I have very recently been trying to come to grips with whether the beta or VBE is what I should be looking at. I do suspect that it may be a combination, but since beta varies more than VBE, it may be the thing we should watch. I don't know.
In fact, I am shortly going to be doing some tests on this very subject. I sure hope I'm right! Either way, I'll learn something and I feel that is worthwhile in of itself. What I can tell you is that I have been matching transistors for many years. I know it makes a huge difference, but by matching beta, I may be automatically matching VBE. I also know that buying in 100 lots greatly enhances my yield of matched devices (by beta). I will bet that matching on VBE would increase my number of matched pairs.
I have attached the jig I use for matching these parts. The method removes the device temperature as a variable as two parts are compared, in contact with each other. A foam cover goes over the two transistors. It works so well that any mismatch in degeneration resistors in the target location makes a big difference. That suggests that I'm on the right track (again, I sure hope so!). I may be barking up the wrong tree here, so correcting my view will cut down on how silly I feel. 🙂
The matcher for PNP or P channel is the same, inverted polarities. Very quickly built, I should use a PCB as I do use it a lot.
-Chris
Attachments
Parasound usually uses 2SK389 and 2SJ109 pairs at the input, on my designs. However, they can hand match 2SK170's and 2SJ74 devices so that they work well. Our biggest problem is heatsinks and heat, not matching. However, the JC-1 output devices have to be reasonably carefully matched, and we had some early problems. In fact, I have a JC-1 right next to me, that I had to go through to keep it from current hogging. I was able to get a good enough match from just a packet of repair parts, to get it to work OK.
As we use servos, input offset is a no-brainer. It is also my position that 2nd and 4th harmonic are trivial problems, EXCEPT for meter readers.
IF I wanted to, I could probably lower the THD of my power amp by 10-20dB. It is just not worth the effort, cost, and implied audio quality reduction, due to the use of a more complex circuit and the use of more negative feedback.
Nelson Pass is really lucky in this regard. He has control of the size of HIS heatsinks, and the performance in the marketplace several of his designs reflect this. I wish I could use more quiescent current, most of the time.
As we use servos, input offset is a no-brainer. It is also my position that 2nd and 4th harmonic are trivial problems, EXCEPT for meter readers.
IF I wanted to, I could probably lower the THD of my power amp by 10-20dB. It is just not worth the effort, cost, and implied audio quality reduction, due to the use of a more complex circuit and the use of more negative feedback.
Nelson Pass is really lucky in this regard. He has control of the size of HIS heatsinks, and the performance in the marketplace several of his designs reflect this. I wish I could use more quiescent current, most of the time.
Hi John,
Thank you sir!
Those dual packages are the only way to stay sane when you are producing many.
You are referring to the size of your heat sinks being determined by the size the marketing department wants to see? At least you don't have the horrible lack of bias control that the guys at Bryston have in their early products.
-Chris
Thank you sir!
Those dual packages are the only way to stay sane when you are producing many.
You are referring to the size of your heat sinks being determined by the size the marketing department wants to see? At least you don't have the horrible lack of bias control that the guys at Bryston have in their early products.
-Chris
I would like for you guys to show your results and differences between simple and complementary differential here. To much words, funny stories, audio novels - quite boring. Anyone can write a post.
Why don't you also do something useful, PMA, rather than just criticizing my distortion measurement capability? Why don't YOU model the H bridge and find the optimum values? You are in the best position to do it right, and it would help us all.
$400 mains cords
If you don't care to ridicule yourself, go ahead, and don't forget the Bybees.Joshua_G said:
scott wurcer said:
Hi Scott,
I've had the same experience with transistors bought at the same time. Mostly like peas in a pod. We do have to test to rule out the occasional exception, however.
Cheers,
Bob
PMA said:
Including you 🙂.
Seriously, PMA, since you were the one who asserted that complementary was superior to unipolar, I think now the burden is on you to post a decent apples-apples comparison. What is your technical reason for why you assert the complementary is superior?
Cheers,
Bob
anatech said:
I just saw the new Bryston 1000W @ 8 ohm amplifier in Montreal with the cover off. It is a bridged design, of course. Each side has four pairs of MJE21193/94 output transistors. This was a bit disappointing, both in terms of total number of devices and non-use of a RET device with ft > 20 MHz. The OnSemi RETs are not that expensive. Power supply reservoir had 16 10,000 uF, 100V caps.
I was hoping they would have moved on to ThermalTraks.
Cheers,
Bob
Bob Cordell said:
Bridged input stages have current on demand and I would think in the end that would win if you want to push out BW. I have never seen a slew enhancement scheme on a unipolar input stage that was very linear. In the normal range of signal level you might be able to get either to produce good results.
Hi Bob,
Well, don't forget they now have a real engineer. I'll bet that there are both time constraints and expenditure constraints he has to deal with. Did they stick with the same "special" output configuration that was their hallmark? It's entirely possible he spends time in the service department. Just guessing here.
I have not really paid much attention to Bryston after I went head to head with them years ago. It was clear to me that many of the claims they made were not possible. For instance, they claimed to match resistors to 0.1% on an AES tour. I looked around and couldn't see any instrument that was capable of even 2%. They seemed to spend more time matching the colour of the heat sinks than anything else. I know there was no output transistor matching.
Interesting you were examining a 1 Kw amplifier. Any idea what their 100 watt amplifiers look like? It could be argued that the 1 Kw amps were not considered for the better devices as sound quality is probably not the selling feature of that amp. What they can do with an amplifier in a reasonable power class would be a better indicator of their abilities I think. What I have heard from others is that they began producing listenable amplifiers with the "SST" series. I haven't had a chance to hear one yet.
Using a ThermalTrak implies an understanding of what they bring to the table. It's interesting because there original 3B/4B amps had defective bias control circuits. Badly under compensated and certainly not engineered. I have been toying with the idea of tearing down one of these and rebuilding it closely. The bias circuit would be redesigned of course. The other two things missing from their older product? Output speaker protection and inrush current limiting.
-Chris
Well, don't forget they now have a real engineer. I'll bet that there are both time constraints and expenditure constraints he has to deal with. Did they stick with the same "special" output configuration that was their hallmark? It's entirely possible he spends time in the service department. Just guessing here.
I have not really paid much attention to Bryston after I went head to head with them years ago. It was clear to me that many of the claims they made were not possible. For instance, they claimed to match resistors to 0.1% on an AES tour. I looked around and couldn't see any instrument that was capable of even 2%. They seemed to spend more time matching the colour of the heat sinks than anything else. I know there was no output transistor matching.
Interesting you were examining a 1 Kw amplifier. Any idea what their 100 watt amplifiers look like? It could be argued that the 1 Kw amps were not considered for the better devices as sound quality is probably not the selling feature of that amp. What they can do with an amplifier in a reasonable power class would be a better indicator of their abilities I think. What I have heard from others is that they began producing listenable amplifiers with the "SST" series. I haven't had a chance to hear one yet.
Using a ThermalTrak implies an understanding of what they bring to the table. It's interesting because there original 3B/4B amps had defective bias control circuits. Badly under compensated and certainly not engineered. I have been toying with the idea of tearing down one of these and rebuilding it closely. The bias circuit would be redesigned of course. The other two things missing from their older product? Output speaker protection and inrush current limiting.
-Chris
Hi Scott,
-Chris
I don't think that is even an issue these days, is it? Music comes nowhere near the slewing limits in modern amplifiers.
-Chris
Re: $400 mains cords
If you think it's ridiculous, please go ahead and have fun.
I do tests before forming an opinion. I find it ridiculous to operate out of beliefs, without testing.
Edmond Stuart said:
If you think it's ridiculous, please go ahead and have fun.
I do tests before forming an opinion. I find it ridiculous to operate out of beliefs, without testing.
Re: Re: $400 mains cords
I do use my brains before testing and wasting my money.
edit: And cut off the wires from the Bybees, so you can use them as it should be. 😀
Joshua_G said:
I do use my brains before testing and wasting my money.
edit: And cut off the wires from the Bybees, so you can use them as it should be. 😀
scott wurcer said:
Hi Scott,
This is a good point, and is particularly relevant to John's JFET complementary input stage. However, the majority of complementary BJT input stages are not bridged and do not enjoy current on demand. Those are the reference against which I am comparing.
Cheers,
Bob
anatech said:
Hi Chris,
They claim it is a modestly improved SST-type design, so I'm guessing they are still using their proprietary output stage, but I was not able to see for sure.
BTW, IIRC, the 4B SST still uses the stacked output stage. I wonder if this one still stacks the output stage.
Cheers,
Bob
- Status
- Not open for further replies.