• The move to the new server is done. There are some software and database maintenance updates in process. This has us passing the hat around to help out. We appreciate any donations. Seriously, even a dollar helps. The payment page may be found here - https://www.audiokarma.org/support.html

DC OFFSET question

If you already have the TIPs use them and start asking those on AK that have some audio experience for modern replacements for your TIPS in TO220 (I use that as a generic for tab style cases). The design is really marginal. You have 67V rails and the TIP 3XC is 100V Vceo. That means if one side shorts you have 134V across a 100V transistor or if you get a bad imbalance you might also exceed the ratings.

From your pictures you have really great bones to work with and can get super amps for your efforts.

If it was me I would be looking at MJW3281A / MJW1302A complementary pair. I do see some questions to be answered, that is a just off the cuff.

I would also look at different drivers like the MJE 15032 /33 pair (also tab type). Even in air without a simple screw on heat sink they are a big improvement, probably cheaper and available almost anywhere.

No dedicated sockets I know of. You said you had a bit of a junk box. Look at some old board connectors that you can hack up to make good 3 across sockets with correct spacing. I would be picky and maybe look at gold flashed connector for no good reason but it would make me feel good

I will look more tomorrow if we have power. Forecast is for really nasty ice storm and prolonged power outages.

The SCRs then would most likely be OVP. Quite strange on an unregulated supply. Also maybe to blow fuses on offset or shorted output. Might want to check the SCR and find if it is blown (probably open by this time) also as if it is offset detection and you had bad outputs it should have popped fuses. As I remember the max amps for it was on the low side for crowbar / fuse blowing if that is it's use in the circuit. A higher rated unit would be the order of the day.
 
Last edited:
Socket

I just noticed the the 3 line folded trace (bottom offset to left of your last attachments) on the board that mounts on the caps. The guys that made these boards were interesting. I don't think I've seen an inductor etched on a board in an audio amp before. Common in UHF and up applications. The only thing I can guess is that the designer was balancing out the collector to heat sink capacitance. Worth while ? no clue? sure would like to hear from the designer or other guesses.

On the driver board. It looks looks the 10 Ohm 1W between the .22 5W and the choke is in parallel with the choke that would be the normal path to a speaker. See a couple of posts back. If you ever have to replace that 10 Ohm stick with carbon or film (no wire wound) and maybe go to 2W if the lead diameter will fit the holes.

TO220testjig_zps08cae3ef.jpg


You ask about a socket. This an end off a cable harvested from an old computer that I made into a test jig for TO220 type. Whack off the extra holes you don't need. If you need longer or different length wires, use a pin extractor or make up a thin metal finger for an extractor, use a dental tool or whatever you can think of to pull the old pins and carefully uncrimp the old wires. I don't have a warm fuzzy feeling about sockets in this application. Your full output current (collector and emitter) flows through those leads. You may be adding a failure point.

I would probably give a socket a go unless an AKer here has had bad experience and says not to. One of my many personality flaws is I love to try things and sometimes dig deep holes to get out of.
 
Last edited:
I soldered in the new output transistors and reconnected the leads to board 2 (protection board?) I power it up on the DLT and no shorts now ,light dims way down and VU meter needle doesn't pined . The amp got down to -87mV off the DLT and then the output voltage jumped to 57 VDC and the 68 Ohm resistor burned. This resistor is connected to the B of the second 2N4410 goes to ground via the 1000uf/10v cap(new cap) ,there are 2 other resistors off this trace that don't burn. The original 68 was the smaller body type so I am using a 1/2 watt
 
Any chance you put the cap in backward? 68 Ohms in series with 1000MF off a base to ground sounds mighty strange. Parallel might make sense, but still a lot of capacitance.

I found a :lmao:FREE:yippy: program mentioned in the December 2013 QST Hints & Kinks section that might be of help.

It is for drawing circuits and supposedly lets you drag the parts around after you put them in and connect them. The author of the hint used it to make a drawing of a controller board from an old yard light. I have not tried it yet.

ExpressSCH from www.expresspcb.com
 
I'll check it out ,thanks .
The cap is in right the 68 ohm comes from the Base goes to the + of the cap and then to the - of the cap then to - on the board. There is no chassis ground that I can see, the - goes to the - side of the input jack but the powered coat on the chassis insulates it. Is this a" floating ground " ?
 
Last edited:
I'm not sure what "DLT" means, but my guess is that you're using a lamp in series with your amplifier's power connection?

That may not save your output devices. You really need to cut off the bias current, if you want to run your output devices unloaded and off of the heatsink.
 
There are not a lot of resistors on this board so I'm going to change them .
The output transistors stay cool when testing the board this way .I'm not sure how to short the bias .
 
Here is Board B "schematic" (output board) .Take a look please . I'm working on Board C (protection board) . Q9 (tip36c) shorts out blows + fuse
Do you think 2n5320/2n5322 would be a good replacement for the 40495/40595 ? if replacements are needed .
 

Attachments

  • img015.jpg
    img015.jpg
    67.3 KB · Views: 18
replacements

The schematic helps a lot. The only difference I noted in the 2 different ones was rotation / not parts or connections.

Is Q9 the only one blowing?
Do you have anything hooked across the speaker leads?

Assuming the speaker connection is still open as in the past posts there should be no path for enough current to blow an output.

Check the ground isolation on the outputs. As long as the other 3 are not shorted, it sounds like one or more or the outputs have the collector grounded. The TIPs come both with and without isolated tabs. If the originals were isolated and the replacements are not you need to add the mica under and screw isolators.

I suspect the coil on the board is across the 10 Ohm 1W connected to the speaker -.

Drivers:
I found another data sheet for the 40595 and its complement 40594 (not 40495) that looks more likely:
http://www.datasheetarchive.com/dl/Scans-005/Scans-00115361.pdf

Looking at that sheet they are listed as a class A predirver in the 50 to 100W range depending on the load. That lists them with the 2N5322 (5323) as the same 40595 and the 40594 as a 2N6320

The data sheet for 2N5322 /3
http://www.datasheetarchive.com/dl/Datasheet-094/DSA0089937.pdf

List that pair as a TO 92 good for 1W. No way would that drive your TIPs bases.

Short form 2N6320 Datasheet lists it as a TO3 300W 140V 25A so there is a major disconnect on the part numbers.

The 40594/5 are listed as quasi complementary pair. Not a good thing these days when better stuff is out there.

I suspect the 50 to 100W is correct as it would fit your application.

Shoot me a picture of the 40594/5 that are actually in your boards. If they are really little TO 92 cases and not a take off on the TO220 style I would guess someone made a bad substitution before you got the amps. Also the pin layout will be a clue of the originals.
 
Drivers & Outputs

All data so far found on 40594 / 95 list them as about 1W. Comset sheet looks like you might get 10W and Ic of 2A with the 2N5320 /22, but no curves.

TIP 35 / 36 have a max Ib of 5A each and gain 25 to 50. That is 10A max Ib for a pair. Actually with a rail of about 60V you are limited to Ib continuous of around 600 ma each from the curves in my 1990 TI Power Products Data book. The 1W driver class to get 100+ WPC output isn't likely to be too happy. The Comset might squeak through but I would not go there.

Since you are sure the 94 / 95 drivers are original the best - GUESS? - I can come up with is the original outputs were Darlington or FET. The TIP 142 / 147 are Darlington with voltage and power rating similar to the 35 / 36.

You may find you are satisfied with the output power as is.

From past posts you have (or had) an oscillation problem. Going from gain of 50 to 1000 may (or not) exasperate tenancy for oscillation. The gain change is what it is going to take to have happy 1W drivers and get 100+ WPC output.

No experience here with audio designs using FET or Darlington outputs. Hopefully someone around AK can chime in with some help. Seems like an interesting design and I will be keeping up to learn something. Looking through the AK DB for amp designs using FET or Darlington outputs would be a good point of departure.
 
There is short somewhere here , it appears to be in the positive rail on the board. When I remove that fuse the DLT dims . I need to set up the board with the output transistors connected with jumpers , I can't keep heating up the board to remove them. Here is the "schematic" of board C
 

Attachments

  • img017.jpg
    img017.jpg
    52.6 KB · Views: 8
I think it might be Q5 (40594) causing the problem . I removed it measured the hfe and its 10 ,the specs have min. 70 . The hfe of the 40595 is 192 . With such low gain would the transistor work correctly ?
 
Short is probably tab to ground on one of the TIPs. Grab a magnifying glass and inspect the Mica insulators.

Depends on what you mean by Properly.

Will it cause fuse blowing like a short no.

Will it cause a major imbalance in the drive to the + half of the cycle to the - half YES.

Distortion YES.

You need the complementary pair of drivers to match as close as you can. One drives the positive half cycle of the output and the other the minus half of the cycle.

Same reason you try to match gains of the NPN and PNP gains of the outputs. This is the reason manufacturers work so hard to make complementary transistors that the output curves are perfect mirrors.

The emitter balance resistors are to insure each transistor on that half cycle shares current. The only thing you have to match the positive and negative half of the cycle is gain matching. The more perfect the gains are mirrors the the more perfect the reproduction is, i.e. fidelity.
 
Last edited:
The TIP outputs are not on the heat sink so the collector is in open air, both sides. The low measured gain is 10 and the spec min. Is 30 .I'm thinking maybe the transistor fails when powered. The DC offset starts to Go toward zero and then it's 50v on speaker terminals and pop goes the + rail fuse
 
You must log in or register to reply here.
Back
Top Bottom