tryonziess said:
The ThermalTrak transistors are a nearly revolutionary improvement in BJTs for audio output stages, in my opinion. Moreover, the ThermalTrak app note barely hints at all of the good things about these devices, and surely sells them short by just emphasizing a reduction in the number of parts in the bias spreader and the possbile elimination of the bias pot. Don't get me wrong, all of these are VERY good things, and they do indeed surely help the non-high-end amplifier manufacturers a lot.
However, for high-end manufacturers, the elimination of a couple of small components and a bias tweak is not where it is at. The thing that is dearest to my heart is the bias accuracy and near-elimination of the thermal lag problem. This allows one to get the bias to be much closer to the theoretical proper Class-AB bias on a dynamic basis, under real program conditions. This greatly reduces dynamic bias instability and dynamic bias shifts away from the optimum. Most people who think they are operating at the Class-AB optimum most of the time with conventional bipolars are fooling themselves, in my humble opinion.
I suspect that dynamic departures from optimal output stage bias due to thermal lag (this is a type of what some call memory distortion) is one thing that is potentially responsible for static bench distortion measurements not always telling the whole story about how an amplifier sounds.
Moreover, the proper bias tracking makes it possible to have a much more reliable bias-setting arrangement that is far less prone to thermal runaway. One way of looking at this is that it makes it safer to use smaller-value emitter ballast resistors, leading to designs that can potentially operate with healthier bias and lower output stage crossover distortion.
Cheers,
Bob
Bob, my main interest in these devices came about when i noticed that Mcintosh uses 96 of them in there benchmark mc2kw amplifier. Possibly being the most advanced amp design they have ever attempted and placing it at the forfront of there dream system it got my interest. Also, i believe On Semi truly is trying to develop exceptional audio components. Having not the time or knowledge, at this time, to build all of the projects i would like. I want to use modern components in my builds. I plan on leaving everything to my grandkids and would like to know that they are worthwhile keepsakes. thanks for you time Tad
Class A
Dear Bob, "Optimum Class AB" is not Class A.
Many prefer Class A for sound quality.....
http://www.diyaudio.com/forums/showthread.php?s=&threadid=104589&highlight=
Bob Cordell said:
Dear Bob, "Optimum Class AB" is not Class A.
Many prefer Class A for sound quality.....
http://www.diyaudio.com/forums/showthread.php?s=&threadid=104589&highlight=
Re: Class A
Yes, I agree. It is easy for us to get tripped up by our semantics. What I meant by optimum class AB was that it is biased in accordance with the principles of Oliver and others that have looked at this. (unfortunately, Self added to the semantic confusion by calling this Class-B). I did not mean to imply that this kind of biasing of Class-AB is as good as Class A.
There is also over-biased Class-AB, called by some Class AAB or the like. This kind of bias gives us a wider Class-A region, but in some cases may lead to so-called gm-doubling, so there is a tradeoff (a tradeoff which I think is usually a good one). While this kind of Class-AB biasing also benefits greatly from ThermalTrak, the so-called optimum class AB biasing benefits relatively more. The reason for this is that the optimum class-AB bias is lower, and there is less margin for error against a serious under-bias situation during thermal mis-tracking.
Cheers,
Bob
QSerraTico_Tico said:
Yes, I agree. It is easy for us to get tripped up by our semantics. What I meant by optimum class AB was that it is biased in accordance with the principles of Oliver and others that have looked at this. (unfortunately, Self added to the semantic confusion by calling this Class-B). I did not mean to imply that this kind of biasing of Class-AB is as good as Class A.
There is also over-biased Class-AB, called by some Class AAB or the like. This kind of bias gives us a wider Class-A region, but in some cases may lead to so-called gm-doubling, so there is a tradeoff (a tradeoff which I think is usually a good one). While this kind of Class-AB biasing also benefits greatly from ThermalTrak, the so-called optimum class AB biasing benefits relatively more. The reason for this is that the optimum class-AB bias is lower, and there is less margin for error against a serious under-bias situation during thermal mis-tracking.
Cheers,
Bob
Bonsai said:
Yeah, I've actually had about 350 in a row. I was hit by an out-of-control motorcycle and now I am paralyzed and in severe constant pain.
So sometimes I overreact.
And you still don't know what you are talking about in this case.
Have you ever talked to Mark Busier, the guy that designed the ThermalTrak parts? Have you ever talked to Todd Visconti, the guy who is the product line manager?
If you had, you would know that you are WAY off base in your accusations. Let me give you just ONE example. Go to the On Semi website and look at the pricing for the MJL1302A. It has a budgetary price of $1.85. Then look at the price of the NJL1302D, which is the ThermalTrak equivalent. It is a whopping $2.39.
A couple of things to note here. First of all, you get a MJL1302A plus a MUR120 diode (that is what is inside), plus you get a 5-pin leadframe, all for an extra $0.55. If that is not a bargain, I don't know what is.
And if that's not good enough for you, just stick with the old version. It's not like they discontinued it. In fact, the budgetary pricing DROPPED $0.15 in the last year, while at the same time they improved the process and the part, tightening up the beta matching by a factor of two or more over the older parts, to say NOTHING of what the "competition" offers. Plus the SOA has increased, plus the Vce has increased (not shown on the data sheet, but there nonetheless).
And Mark Busier IS trying to give us better parts to build better amplifiers. Of course On Semi is hoping to make money by doing so. What did you think, that they were running a charity?
Bonsai said:
Yes, I bet you do. It sounds like you work for a competitor that makes an inferior part and feel the need to badmouth On Semi instead of recognizing that they have built the EXACT part that knowledgeable audio engineers have been wishing for for decades.
And it also sounds like your company operates under that cynical attitude where profit is all that counts, and the bean counters are in charge, and you project that cynical attitude onto everyone else.
My condolences to you.
If you like what you are doing, you have my sympathies. And if you don't like what you are doing, why don't you see if they have any openings at On Semi. Those are great guys building great products that let people like me build great amplifiers -- much better than what I could build if I had to rely on the products made by their competitors.
So next time, don't go around bad-mouthing Mark Busier, Todd Visconti, or On Semi. Some of us here know those guys, have worked with them closely and respect them highly.
Charles Hansen said:
Hi Charles,
A little OT, but I wonder if these guys are aware of the problems with the SPICE models they publish for these parts? I've got some pages on my web site that provide many specifics, and improved models for them. Could you email me with contact info for them? If so, I'd appreciate it.
Apparently they subcontract the model extraction stuff out, and whoever is doing it does not check the end results. That person seems to just plug the data into a program called MODPEX that in some cases spits out completely wrong parameters.
Charles,
I am sorry about your personal situation - I hope things improve for you.
I did not mean to foulmouth On - just an observation about the realities of the semi indistry.
I work in the industry but not as a direct competitor of theirs.
I look forward to seeing some practical DIY Thermaltrack designs on the forum.
I am sorry about your personal situation - I hope things improve for you.
I did not mean to foulmouth On - just an observation about the realities of the semi indistry.
I work in the industry but not as a direct competitor of theirs.
I look forward to seeing some practical DIY Thermaltrack designs on the forum.
Digikey is very convenient, but you pay a price for that convenience. Mouser is often cheaper on many items.
But if you are going to buy $50 or $100 worth of an item, try a real distributor. Arrow (www.arrownac.com) has the non-lead-free version listed at $2.36. I don't know what the minimum order is or the handling charge, but check it out. You probably want to buy a couple of dozen just to play around with, so this route makes a lot of sense.
But if you are going to buy $50 or $100 worth of an item, try a real distributor. Arrow (www.arrownac.com) has the non-lead-free version listed at $2.36. I don't know what the minimum order is or the handling charge, but check it out. You probably want to buy a couple of dozen just to play around with, so this route makes a lot of sense.
Charles Hansen said:
Charles,
This is a good point. For samples, however, I got five "samples" of each directly from onsemi (off their web page) for about $11 postage and handling. Probably just a one-time thing, though, assuming they care and keep track of who is requesting samples.
Cheers,
Bob
Hi Bob,
It is best to treat the samples program with respect. Some people don't and that's why you can only get 5 pcs now.
I do have a great deal of respect of On Semi and their parts. BTW, Mouser is terrible to work with from Canada. I tried and it was ugly. They only use UPS. Digikey and Newark are were I buy larger quantities once I have figured out what I need for parts. They normally do have some quantity breaks. Some times that is a rail, or 25 pcs per number.
-Chris
It is best to treat the samples program with respect. Some people don't and that's why you can only get 5 pcs now.
I do have a great deal of respect of On Semi and their parts. BTW, Mouser is terrible to work with from Canada. I tried and it was ugly. They only use UPS. Digikey and Newark are were I buy larger quantities once I have figured out what I need for parts. They normally do have some quantity breaks. Some times that is a rail, or 25 pcs per number.
-Chris
Charles Hansen said:
Makes the reading a lot more colorfull though.
A lot of people forget or have never known how affordable semiconductors have become, relatively speaking.
Same goes for shipping charges, i'm still frequently surprised how little it costs to have bulky stuff transported across thousands of miles distances.
Charles, Thanks for the Arrow electronics tip. I will see what they have to offer. By the way do you think the thermal trak would couple up to the National LME69810 very well in a system. I do not know if the driver has a way to accept the diode circuit. I guess i will have to wait for someone to post a design. Thanks Tad
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.