Dx Troyan, a 650 watts channel amplifier.

Status
Not open for further replies.
here is my PM to DX in it's entirety. I cannot understand why DX cannot read it.

VI limiter
Do you think the V/I limiter is reasonable?
It is working and limiting....i could not find failures till this moment....do you have suggestion to make it better?...............................
Can you help please AndrewT?
Yes.
But, the component values in the VI limiter depend on
Rth s-a
Rth c-s
output device SOA
output device temperature de-rating
Ta
Tc
transformer voltage
transformer regulation
smoothing capacitance
output emitter resistor value
matching of output devices if more than 1pair
matching of Re if more than 1pair

I have an Excel spreadsheet with a VI limiter built in.
You will see that altering just one of the above variables can throw the VI limit locus way off from where the builder might want it.
Oh, that's the other variable.
Where does the builder want the VI locus compared to the DC and 100ms and 10ms SOA limits?

Do you want me to send you the spreadsheet? It is based on Bensen's FET SOA sheet with many modifications since he posted it years ago.
The model that can be built up in the spreadsheet turns out to be quite accurate for current and voltage into resistive loads and allows one to see the effect of reactive loads.
One part that I cannot model is transient currents into reactive loads. I have to guess at that bit and use experience and some actual testing.

Here is my philosophy on a VI limiter.

The limiter must pass unaltered, all valid audio signals to all valid speaker loads that are within the specifications of the amplifier.

It should hopefully protect against damage when the amplifier is abused.
If the VI limiter meets that specification then it cannot affect the audio quality of the signal passing through the amplifier.
regards
Andrew T.
__________________
regards Andrew T.
 
Administrator
Joined 2004
Paid Member
Hi Andrew,
I cannot understand why DX cannot read it.
Please cut Carlos some slack. You know that English is not his first language.

As for "the amp will not work" type comments, they are not accurate. A more accurate way to get across your concerns would be that running with a single pair of outputs into a 4 ohm load isn't recommended. The amp will work, but it may fail when running at high level into a four ohm load. The amp will work after all. It just may fail.

For years I've seen amplifier repairs that shouldn't have survived turn-on. I've learned that electronic parts can sometimes put up with all kinds of abuse. So while running parts beyond their safe operating area should be discouraged, the parts may just survive in spite of what the spec sheets say. Remember, the manufacturer states that they will not guarantee the parts will perform beyond the limits they give. That doesn't mean that they will not survive.

Does that help?

-Chris
 
Why? Either the peak junction temps don't really get that high (even in the s/b limited region), or the peak junction temps can be run higher than any 200C. In either case, the semiconductor makers are sandbagging their specs on the modern devices.

Hi! How's the amplifier building coming along? :)

I think you are right. Or more likely, it's both those factors combined. The thermal capacity of the transistors is NOT negligible. Temperature variation over a cycle isn't really that great, even at 20 Hz.

As for second breakdown issuses, the "dangerous" part of the dissipation pulses when driving reactive loads is about 10ms in length even at the lowest of frequencies and the duty cycle is low. In such a situation, transgressing the 100ms S/B a little is not going to be fatal. +-100V rail class A/B with 2SC5200 style devices sounds a bit excessive though. Most manufacturers seem to start using series output stages (like the Super Leach) above 80V or so.

There also seems to be quite a bit of conservativeness in the temperature ratings. I was dissapointed that I couldn't blow up a 2SK176 (a TO3 lateral MOSFET rated 125W @ 25 deg C) with a power dissipation of 200W mounted to a heatsink kept at 90 degrees C! I gave up after about a minute. Tough little guys these transistors. ;) Sure, a BJT isn't as internally themally stable, but instant destruction sure doesn't happen at 200 degrees C like most here seem to think. A Sanken note in a databook suggested it's more like 260-300 degrees C and that sounds about right... :hot: The LM3886 "SPIKE" protection activates at 250 deg C power transistor Tj if I remember correctly.
 
Administrator
Joined 2004
Paid Member
Hi Pavel,
Thank you. It's really good to post a bit.

I agree Andrew is trying to help, but I'm afraid that our two combatants may be misinterpreting the language. My post was more to try and have Andrew and Carlos step back a bit and not pay so much attention to the details. From what I had read, Carlos has a valid argument in that he is running those amplifiers now. I merely suggested that the reason why this is so is because of "safety margins" built into transistors.

Hi Andrew,
No, I have not seen the Vas area. But I'll believe you that the transistors may be exposed to excessive C-E voltages. However my intent and post was more about about allowing some room in language. Note I did not say you were wrong. What I said was that it is entirely possible that Carlos has some of these running merrily away with self destructing right away. My only problem was the phrase "can't work" as opposed to the more correct idea that the amps can work that way. The fact that it is not recommended to run transistors (or any parts) beyond their maximum limits and they may fail in operation is another thing again. I think that this is what you were thinking when you posted, but the word "can't" is the wrong word. "Shouldn't" seems to more accurately reflect the situation.

Ahhh, when words get in the way ...

It was posted just a short while ago that the 818 was only in simulation.
Post 818? What are you referring to there Andrew? I just sailed in after receiving a reported post.

-Chris
 
Post 818? What are you referring to there Andrew? I just sailed in after receiving a reported post.

-Chris


A BC818 that is connected in a way, driver of the
darlington Vas , that it have to sustain a full rail voltage,
that is , 100V at the start , and currently 80V since the
PS voltage had been reduced.
Anyway, surely that he did use something else when
building this amp.
 
AX tech editor
Joined 2002
Paid Member
It doesn't surprise me a bit that it works - Uncle Charlie isn't taking any more liberties with SOA than QSC, Crown, and everyone who copies them haven't been taking for years and years and years. Less, actually. Look at some of those designs and you wonder why they don't just blow up the very first time the bass drops. Instead, the only thing that seems to kill them is long term heating.

Why? Either the peak junction temps don't really get that high (even in the s/b limited region), or the peak junction temps can be run higher than any 200C. In either case, the semiconductor makers are sandbagging their specs on the modern devices.

Correct, but the thing that makes all the difference is that QSC, Crown etc have sophisticated protection systems, while the Troyan only has an I/limiter which is not sufficient at these voltage levels.
Also note that a lot of really high-power amps have 8, 10 or even 12 pairs of output devices.

That is really the only argument here: make sure that you have enough devices to stay outside the danger zone, or put in a good protection system.
Sure the Troyan 'will play', will survive full power in a 2 ohms resistor - that's also what AndrewT found and clearly stated.

But hang a real speaker on it and the game changes.
Many people find that hard to believe because they are unfamiliar with these issues, but anyone who has used these types of amps and has been plagued by seemingly unexplainable random failures even when not playing full power knows.

And if you have an output stage failure in a 8-pair output it's a hell of a job to find the problem, let me tell you that. So if you build this amp, build it first with a single output pair and test it with a higher load resistor to make sure it is wired correctly and it will survive the full output voltage level. Then add devices pair by pair, and re-adjust the bias and recheck everytime. Amp building is fun, amp fault finding isn't. YMMV.

jd
 
Anatech,
go to post80.
I asked 4 technical questions to try to understand the philosophy behind the schematic.
To date no one has tried to explain why those technical details were adopted, or at least I cannot recall seeing any answers.

It is preposterous to specify a 25Vce0 device for a 100V location. See post11 & post15.
It is irresponsible to recommend this schematic with this advice to every Member who chooses to follow Destroyer's teachings.

That to me is wrong.
It is not opinions that we are discussing, it is not shades of grey, it is not design decisions based on alternative briefs. Post189 confirms my belief that differences in opinion can be discussed technically. We may not agree but other targets can co-exist.

Post190 is the first occasion that I can recall that destroyer told us that the bc818 was only in simulation and that some other component was used in testing. He has still not told what he tested. Yet from his first post he recommends to his Followers this 1000W proposal.

Destroyer specifically asked for me by name to help with the VI limiter. I sent him a personal message (PM) explaining what I could do and asked if he wanted a copy of the modeling software. I got no reply. Later Destroyer posts that I choose not to help.
Then he says he cannot read my PM to him. Tell me why he cannot read a PM from me. Is it possible that he has blocked PM as well as posts from me.
Then why ask for my help?

This is not a language or interpretation issue. Destroyer asked me and I responded by offering to help.
What more do you expect?
That to me is quite clear. I did not do any wrong. I did not manoeuvre Destroyer into an uncomfortable stance. He did that to himself.

Now, I'd like to suggest the Moderating Team go and read all the safety warnings that various Members have repeatedly given in this thread. Read what Destroyer's responses to this advice was. Decide if this thread is safe to leave on the Forum. If the Moderators leave this thread open, then I challenge you to insert some safety advice so that Members do not come to any harm or financial loss.
Without some form of cautionary statement, the Moderators will be failing in their obligations.
 
Last edited:
diyAudio Chief Moderator
Joined 2002
Paid Member
Is there any recommendation in that project by the author that involves human exposure to danger beyond typical high power amp assembly practice? Did you make that project and found out that it is highly dangerous, more than any average high power amp project? Or is there any sure safety rules violation for powering it? You used the term ''harm''. That's serious. We did not have any safety danger reports yet.
Give the links to the posts that alert to sure serious dangers so we review. If you exaggerate and you are calling out the moderation team for will be failing obligations on such a serious matter in the open, by bypassing the related rule ''Please handle questions and comments about how to improve the site or moderation via the "Report Post" button, PM, or the contact form.'', we will also have to review your post against discrediting another member's work without hard evidence, and for dragging the moderation team by spreading heavy warnings without consulting the team first.

Come up with the links please. If there is ground we will post a warning.

:cop:
 
I think there is a certain level of people having to exercise their own caution when finding information on the net - don't believe everything you read and don't expect things to work on "wishful thinking". The blatant error is the BC818 and any amp would simply blow up or be a time bomb if that device was ever used. Carlos should have corrected it quickly.

The economy in design will really hurt (be catastrophic) under short circuit conditions (where the devices see large V and I) without a good VI limiter, and it must monitor V because it is the combination that kills devices. It is also reasonable at this power level to not allow the amp to run at full (worse 1/3 power) on the bench continuously since this requires tremendous heat sinking, it could be spec'ed this way and protected by a thermal cutout on the heat sink. However, considering the high level of compression used in music these days and that max dissipation is at 1/3 max power it is possible to overheat such an amp in demanding pro applications. Fans and proper thermal analysis is the obvious answer with a thermal cutout as further protection.

So, Carlos could simply say that the amp is rated for so many watts into say 4 and 8 ohms and not recommended for lower impedances. He could state that it is 8 ohms only for highly reactive loads. There are some speakers that are essentially a resistive load after all. We all know that people do not follow specs, and the VI limiter should be designed to protect the amp with out of spec loads.

Plenty of amps are known for reliability issues and this will just go down as another one if precautions are not taken. I suppose that there is a fire hazard when devices are used at 3-4 times their rated spec where it might cause the entire amp to blow up and this would be a good reason for the mods to get involved here. It is a no brainer guys ...
 
Last edited:
diyAudio Chief Moderator
Joined 2002
Paid Member
We get involved in special matters only after reported posts. That was bypassed and its a rule. Since there is open talk posted about safety though, we are expecting Andrew's links so to have one or two of our amplification experienced mod team members to review. Thanks for your additional technical input. :cop:
 
I think Andrew's main concern here, is that a 1000 watt amplifier isn't a trivial beast and involves far more potential for danger, then a significantly lower powered amp.

This in itself wouldn't be much of a problem, but there are people who are reading this thread who are enthusiastic towards the design and are intending on building or are at least showing interest towards trying this out. Comments have been posted by suitably qualified individuals towards the safety of the design and these comments are seemingly brushed under the carpet as it is something they don't want to hear.

This is worrying.

The amplifier isn't anywhere near ready for anyone to try out. It has to be unequivocally safe and 100% stable into any load that would be connected to it. And indeed any load that shouldn't be, but often are.

I have been working on a higher power version of the blameless amplifier that I have built and I am extremely reluctant about posting the final schematics due to safety concerns or rather being 100% sure that the protection system is suitably designed. This is after I have spent many many hours pouring over power dissipation spread sheets, data sheets, SOA curves, V/I limiter spreadsheets and having also done many hours of simulation around the protection circuit.

I am now in the process of testing the amplifier in question, to get all the bugs ironed out of it. And even when those are smoothed out, and the protection is working and triggering correctly, I will still be hesistant about posting the final schematic.

This is something that should most certainly be done with an amplifier such as the Troyan, but quite obviously hasn't been.

Douglas Self states himself that the primary concern, that should ALWAYS be the number one criteria that you design around is - thou shall not kill. This is the same for any part of DIYaudio. Now the design problems people are having with this amplifier probably wont result in anyone's death, but they are closer to killing someone then they should be.

No one is taking any personal stabs towards Carlos, indeed everyone here is simply trying to help him. But certain actions need to be taken towards this design to make sure that it isn't a liability.

Simply having large smoothing caps connected up to 100volts in itself is dangerous. If ample fusing isn't installed in suitable locations one could cause an explosion if they are not careful.
 
Simply having large smoothing caps connected up to 100volts in itself is dangerous. If ample fusing isn't installed in suitable locations one could cause an explosion if they are not careful.

It does happen ...

Linn Inc. Recall of Power Amplifiers for Sound Systems

(poor derating of PS caps)

Even to the tune of 2300 amplifiers... :no: :eek:

Amplifiers Recalled by Krell Industries Recalled Due to Fire Hazard

(a simple fuse of the wrong value !)

With the internet we can learn from the misfortune of others.
OS
 
diyAudio Chief Moderator
Joined 2002
Paid Member
PM received. Containing no less than 36 (!) posts to review, it will be taking long enough until there will be spare time for one or two mods knowledgeable to fully see about what order of safety problems we talk about. Therefore this thread is suspended until further notice. By then maybe DX can be expiring his sentence and show more data from the working amp too. Thanks for now.:cop:
 
diyAudio Chief Moderator
Joined 2002
Paid Member
We looked in to it thoroughly and we will not go about issuing any review or warning about any thread specifically. Only for illegal mains AC practice reports we will close threads as we have done before.

As with any project involving high voltages and high power, caution is well advised. There are sticky threads about general AC safety and high DC voltages at the top of forums also.

DIYA produced a disclaimer given this opportunity, soon to be added in the rules section footnotes too.

''Posts and projects are those of individual members of diyAudio. The forum itself is not in the business of vetting projects or posts for safety, accuracy, performance, reliability, function, or fitness for use. If you attempt to make something and it blows up, or turns expensive parts into charcoal, or just doesn't work the way you were hoping, that's between you and the person posting the project or idea. The forum is merely a bulletin board which allows anyone to post ideas, criticisms, or discussions. You have to decide for yourself what makes sense and what doesn't.''

Therefore the thread is opening again. Use the report post button only for any matter rules & moderation, refrain from personal remarks.

Regards

Mods team.

:cop:
 
Status
Not open for further replies.