Power amp under development

quasi said:
Hi,

First of all, what are your voltage rails?

2x50V, from an adjustable voltage, adjustable max. current powersupply.
(I could call it "laboratory PSU, but it's not a lab. PSU 🙂 )

If the heatsink of the of the MJExxx is cold it means you have no (or too little) current flowing through here.

You should check the components and for continuity for all of the second stage.

Thanks, I gonna check it.

Are you using my PCB layout or did you do your own?

I made my own "version".
 
Hi Everbody,

I had just draw this schema, based on Quasi's output idea.
I wouldn't create a high-and, state-of-art design, only a simle and cheap Public-Adress amplifier.
IR's IRFP240 is relative cheap (about 1,5Euro/device) here, and 200W would be exactly good.
On the weekend I'm going to bulid it, but before I start it, I would hear some feedbacks.
So, what are you thiking? Every opinion is really welcomed.
Lot thanks, best regards,
Edl

(I don't care about unsimmetric clipping, I wouldn't use higher rails for the driver, because I want keep things simple. As I saw, Quasi's last circut neither uses higher driver rail.)
 

Attachments

  • quasicomp_1.gif
    quasicomp_1.gif
    14.2 KB · Views: 2,865
Case and heatsink ready...

Hi Quasi,

I bought case and heatsink, I'm just waiting for your comments on the second amp before I decide which amp to build...

Here is the pic...

Regards😀 😀 😀
 

Attachments

  • case_heatsink.jpg
    case_heatsink.jpg
    60.2 KB · Views: 2,342
Your hFE Experiences

Hi Quasi and others,

I am still working on sourcing the transistors for my "flavour" of your amp. I like to know what hFE experiences others have had based on my questions below:

1) For the 2SC1845 what hFE code (P, F, E, U) did you use? If you know or still have left over spares what hFE do a few of them measure? Any idea based on the part who made the part and how you were able to tell?

2) For the BC546 what hFE code (A, B, C) did you use. If you know or still have left over spares what hFE do a few of them measure? Any idea based on the part who made the part and how you were able to tell?

3) Many of these small TO-92 parts seem to have a very small space for any info marked on them. Some of the BC parts have "PH" in the black plastic case colour on them via a light grey background for entire flat side. Would the "PH" mean Phillips?

I like to know as I have found a few "interesting" results in measuring various devices and samples Some of the BC devices, the BC546C excepted, do not seem to be able to measure the hFE. The meter in these cases just shows "1" for the hFE with some showing a value like a hFE before for split subsecond then goes to 1 or zero depending what meter I use to measure the hFE. I have been measuring the hFE in part to validate the hFE code of part and/or manufactuer.

Thanks in advance for your information and reply. It will help me so much to sort out a few of your experiences adn what you used.


Regards,

John L. Males
Willowdale, Ontario
Canada
16 March 2006 05:58
16 March 2006 06:04 Typo corrections again, sorry.
 
Hi,
I'm wondering if your hFE measurement is going out of range.

Try going back to first prinicples using a good transistor and measure it the long handed way with resistors in the collector and base leads to allow you to monitor the currents through each.

High base resistor will be affected by the DMM resistance but hopefully not too much.

Two DMMs will make this slightly easier and certainly faster.

When you get repeatable results then try a few of the odd ones.
 
Hi John,
How are you trying to measure the gain? Some inexpensive meters with an Hfe measurement only work over a limited range. Pinout can be a problem too. Pro-Electron parts may have different pinouts than Jan or Japanese parts.

-Chris
 
Hi Andrew and Chris,

Thanks for both of your replies.

Andrew, my suspect sense is you are correct that the hFE I am measuring may be going out of range. Why I say may is of course from the "1" that displays with no demical point. On other hand while inserting the device a value pops on the display for subsecond moment that looks like it might be the hFE well withing "range" then "not" in range. I used the meter I had to do some testing, as well as a couple meters in the store that are "more", but may not be any better in as far as the hFE test.

Having said that I am obtaining very repeatable results for many different parts and samples. All devices I try of same part do in fact behave the same, as in readings or not for a given part. Many are giving a hFE reading. Many are showing a hFE reading as expected based on the part specs.

The reason I asked specifically for quasi's reading for the 2SC1845 is the ones I have tested are supposed for be the "E" hFE rating as marked in the device. An "E" rating is supposed to be a hFE of 400-800 per the spec. What I get for the 10+ parts I sampled is 170-220. Ergo my question so I can compare with quasi's couple of checks, assuming quasi has some spares left over. (quasi have spares 😉 ) On the other hand, the few 2SC2240BL's I sampled had an hFE reading of >425. The BL hFE rating is 350-700, therefore within expecations. Oddly some specifc BC4xx/5xx parts had problems with a measurements, yet some others while others like a BC546C had a hFE of >600. The BC560C's I tested posed a "1" reading challenge. This is why I asked quasi about the 546? I thought quasi still had so much "home stock" 546. 😉. I assumed, though should of stated, quasi would use a hFE test on a multimeter rather than a special hFE meter or dedicated hFE test circuit. I am assuming most multimeters use the same hFE testing circuit. If that assumption was correct I could compare my result notes with quasi's. These may have been incorrect assumptions on my part.

It is my intent to build a hFE tester for more accurate measurement. The primary purpose would to do a simple method of matching. The downside is use of such a tester is not too practical for "sanity" checking" a few samples of a device in a store before I make a purchase.

The challenges I had are not due to pinout differences/options. The sockets for the testers I tried, as well as the meter I have, have sockets layouts that handled all the pin outs combinations I have needed to test.

I had a very good dedicated capacitor meter I have taken with me the odd time to test a few samples of some capacitors of interest before selecting which ones merit purchase for more testing at home. I am not aware of a hFE tester, beyond those on many multimeters, to use for "sanity" checking of perspective devices. I do have a couple of very HP good handheld meters, but neither has any hFE testing. If such a tester exists it needs to be of a reasonable price for the casual useage frequency I have.

I do not have the test circuits handy I found so far searching on the net some months ago. I seem to recall Rod Elliot had a hFE circuit tester, but not really practical for store "sanity" checking.

Andrew, I do not recall seeing any hFE test circuits that used two meters. I do have some very good bench meters I can use for a two meter method at home. Do you know of any links/articles I could read/look at to see how hFE testing with two meters is done?

I do intend to do more searching for a home based hFE tester once I have some more time to do so.

If either of you (Andrew or Chris) have suggestions to based on my orignal question or above comments I am all ears.


Regards,

John L. Males
Willowdale, Ontario
Canada
16 March 2006 20:57
16 March 2006 21:02 Typo corrections despite my proofreading efforts
 
Hi John,
As a quick test I often use my Heatkit IT-18 transistor checker. It is an analog meter type. Not terribly accurate for high gains and it does not work with darlington types. It does give you a good idea where the gain figure is.

As an added bonus, it measures leakage in two modes. I use that a lot in repair work.

-Chris
 
Hi,
the standard hFE measurement uses collector and base resistors.
You need to measure the current in each. You measure the current by measuring the voltage across each resistor.

You can do this with one DMM but you have to check both voltages often trying to set one to a preset value.

This becomes much easier if your plugboard has the tappings for two DMMs and you can use a variable resistor in series with the measurement resistor to set your desired test current and immediately take your matching voltage off the other DMM.

Saves so much time and easy to set up.

If I were a retailer, I would chase you from the shop if you dared to suggest that you select transistors by measuring the hFE before you will agree to purchase.

Buy a batch and use best matching pairs for selected sensitive areas of your design and keep the unmatched for less critical areas of the circuit.
 
The matching world according to quasi;

Hi Folks,

The transistor matching world according to quasi goes something like, don't bother, just use it. In theory matching the 2SC1845's will improve the linearity of the stage, but I doubt it could be heard or even measured, unless there was a huge difference. Like wise the current mirror transistors could be matched etc.

IMHO if someone subscribes to amp topologies that use global negative feedback (like most A/B designs) for distortion achievements of 0.0n% then they have already crossed too much sacred audio ground to worry about matching. I do match output devices if emmitter resistors aren't used, but this is only to try and achieve a degree of current sharing.

I guess some readers will be trying to climb back onto their seats now.....so I apologise for any bruises.

Ok...back to the point. In 2SC1845's I have 14 P's, 2 F's and 1 E left. I have no idea what I used in my modules and have no way of finding out in the near future (audiophile friend is proving to be elusive in the return of said loaned amplifier).

I have checked my batch of 2SC1845's using the Hfe test on my digital meter and the E came in at 390 and the F's came in at 316 & 339. The P's had a spread from 253 to 269 including a surprising 6 units all at 266. The test is probably not that accurate in absolute terms but in relative terms it is good enough.

I would not be concerned about the actual gain though. The total gain of the amp is tied to 33 in any case. I have seen designs that specify MPSA42's and these have a gain of about 80 (from memory only).

As far as the BC546's go, again I don't know what I used but in the ccs application it will not make any audible (or measurable?) difference.

Other stuff;

John I don't know why your E batch are only coming in at 170-220. Are they genuine NEC?

Yes I only use my multimeter to check the Hfe's of small signal transistors, because I only care about the relative result not the absolute. I do have a seperate test jig to check power transistors and FETs for the reason outlined above.


Anyway...avagreatweekend.

Cheers
 
Hi John,
The Rod Elliot design is far more complicated than the IT-18.

I have confirmed the IT-18's accuracy with power transistors by measuring separately Ib and Ic in a jig I built up.

Quasi,
I use a DVM Hfe function for signal transistors as well. I fully agree it's the relative rather than the numeric gain that is important. Unlike you, I have seen differences in matching pairs in certain locations if the amp design is up to it. Strangely, customers have commented later without being informed.

-Chris
 
AndrewT said:
If I were a retailer, I would chase you from the shop if you dared to suggest that you select transistors by measuring the hFE before you will agree to purchase.

Buy a batch and use best matching pairs for selected sensitive areas of your design and keep the unmatched for less critical areas of the circuit.

Hi Andrew,

As I stated previously, I am just testing a handful at best in the store so I have a sense of either what hFE grade and/or manufactuer of the device is. Case in point I bought all (72) of the 2SC2240BL's once I spot checked 3-4 of them and also eyballed rest as I put each into the antistatic bag I brought with me. Same applies to the various part numbers others I have tested. In fact I tested many just to doucment what the store had so I could then review my notes against teh pec sheets and what is likely to suit my needs.

As I mentioned it was just a sanity hFE check or to confirm to ID the devices or their hFE "code".

The more exact and where critical hFE matching is required is in fact to be done at home for each device part after buying more than need to allow for nornal lot differences.

Thanks for the information on how to do a two meter measurement for hFE. I think that will be most useful when time comes to do more critical matching.


Regards,

John L. Males
Willowdale, Ontario
Canada
17 March 2006 20:55
 
Quasi,

I am not brused by you comments on hFE matching or like. I am a "big boy" quasi, I can handle it. 😉 *starts to cry* lol lol sorry quasi I am in silly frame of mind.... I just thought seeing as I have a "batch" of a part, why not take a bit of time and measure the hFE and use ones clost to same value in their "LTP's or mirrored sets either side of rail of driver stage. Hey this is DIY, maybe one will not hear the difference, but at least as a DIY project for me I do not mind taking the time to measure them out and match them up.

As you and Chris mentioned, I am not concerned about the absolute hFE value per se. I would match based on the relative value using the same meter for all devices for a given part number. That said quasi, your question of am I sure I have "E" hFE group or are they are NEC devices or not is the root reason I am measuring a few samples in the store of many of the devices of interest to maybe help me answer such questions. Given the TO-92 form factor and in many cases it seems manufactuer do not even enough space to put full part number. As a result the "S" (or whatever the first letter is) is often dropped. I cannot say if the 2SC1845's I bought are NEC or not.

I took a close look ad the 2SC1845's I purchased. It has two lines of lettering in a very light grey very thin stroke. "C1845" and "E $F" The lettering is on the flat side of the TO-92 case and runs from lead end of case to top of case. Are the 2SC1845's you still have as "sapres" marked in similar manner quasi? I ooked at the top and rounded side of the 2SC1845's I have and there is no otehr markings, therefore no idea if NEC or not.

Quasi, as FYI I was not concerned what you did nor not use in the amps you built. I am sorry if my choice of words suggested that. I was just curious about what might be the values in the quasi "spares store(s)". I do not know how many store outlets you have 😉 Sorry quasi I could nto resist. It has been long hard day with a electronic manufactuer sending the company several hundred skids of completed product with the same serial number they already shipped few weeks ago. long story you do nto want to hear about. Sorry for the digression there.

My question regarding the BC546 for you quasi was just to know what hFE "flavour" you had in the quasi store and what the hFE of the meter said was the hFE.

Ergo how well the meter you, Chris, and I hsed to measure the hFE. I was hoping you would just have the simple hFE provided by a multipurpose digital meter. An as you have commented, it seems that is case *pun intended*.

quasi, Chris and Andrew, thanks for your thoughts, ideas and comments. quasi I used lots of ice so I have no brusing 😉


Regards,

John L. Males
Willowdale, Ontario
Canada
17 March 2006 21:43
 
The Quasi store ....

The Quasi store (1), actually shed at rear of yard, has 73 BC546's and 72 of these are Motorola BC546B's. Number 73 is a Phillips.

This revelation at least tells me what I probably used in my amp.

Anyway I measured a selection of 30 and the Hfe spread was 198 to 235....hope that doesn't cause any tears 🙂. In my amp the only reason I used a BC546 was to cater for a wider range of resistor for R6 and a wider range of voltage rails. Anything similar can be used here provided the Vce is not exceeded. A BC550C would be perfect.

Regarding bruises...lol...my height and ability does not bruise others easily....AKSA the only DIY member who has seen me in the flesh is relatively (maybe absolutely) a giant when standing next to Quasi.....I'm glad there are significant oceans between Australia and Canada...lol!

Hey Chris,

In the circumstances where your client noticed a difference did you match a new transistor to an existing one or did you change both (all). If you did change both (all) did you use the same type or manufacturer or something better / worse (matched). You get the idea ....I hope.


Cheers