My first amp

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Good luck... I also had a cranky new build today... first made a stupid mistake in the vas... then fixed it and got it to play, but forgot the rail resistors in which starts a poping distortion after some time if driven...
Took them out and looked at 56VDC outputs...must have been some "metal dust" as I desolderd all transistors to check and found nothing wrong, powered it on, and it was all fixed again...

I still clearly remember circling my first completed DX amp for hours before switching on... think I took about 5 minutes to blow a bd139....replaced it, then realised I put the vas transistor in inverted, because it poped like a gunshot.... pure ignorance... but it is a good school.

Once you can build and faultrace your DX, you have masterd the majority of old solidstate gear...

MORE photo's please...
I've finished construction! Will put photos up, once I'm home, this is from uni.

Well.. not quite done, havent mounted my second amp on my heatsink.. and the first amp is really shoddily mounted; my dad and I arent the greatest of handy men; you'll see in the photos.

I'll get my transformer today, test later tonight once I get home. A couple final things. Earthing; the ZE point on the schematic is the Earth point, through that big res and the 100nf cap; do I just follow standard earthing procedures with that. I've left out my fuses on the PS; I'll probably install them at a later stage, just the board is too small to fit them yet; once I have the chassis, I'll see how best to install them; The PS should run without the fuses? Additionally, I soldered those 100nf in parrallel with the 4.7mf caps on the PS to the underside of the board; they're small ceramics and they lie flat against the board; there was no space on top, that should be fine?

Insulation; my power transistor insulation kits are weird; the transparent insulating sheath fits the transistor, except for the top two corners, which have triangles cut out of them, so the whole metal surface isnt insulated... pretty weird? Anyway, any ideas for insulation otherwise; would transparency sheets cut appropriately work well? I suppose I can buy another kit, hopefully of the right size this time.

Anyway, I'll put up pictures tonight, might make a little googlepages site and dump some pics on there or something.
Hi the cap and resistor is a common old zobel filter, and connects the output to ground... it forms a filter that is supposed to drain high frequency (dangerous) output to ground.... you can simply run a piece of wire under the board from it to the ground star...

I would not worry about the corners... the rets still forms a spacer... just check for no continuity between centre pin and heatsink after installation....

To get the amp going, you run without fuses, but with power resistors installed in their place.. the ones you will use to set the idleing current.

When it runs you should swap them for fuses.

You did the right thing with the small caps.... the smaller the trace to the thing they are bypassing, the better...
Aaaaaah... Burnt my room down!!

I lie... but i did burn two resistors out. I started testing, and then burnt the 3.3 on the negative PS rail, with the induction coil, and a little more importantly, the 82 ohm resistor on the V- rail on the amp.
Of course, everything needs a story, so here goes. Set up my first built channel to test and set biasing; so my 10ohm, 10watt res's in the rail. Anyway, I turned it on, PS fired up (LED's lit), and I got a voltage across both of those big watt resistors, of +-300mv... Which is remarkably lower than what is needed. Anyway, no amount of adjusting VR3 could get them anywhere more than 350mv... instead of the required 800mv. Nevertheless I decided to continue (bit rash, i know, my bad :/) and have an audio test. Brought out some old speakers, 8ohm, hooked the positive speaker terminal to the amplifier output, the negative to ground (presume that is right?) I then untied my inputs, and put the respective signals into them, playing off my ipod for test purposes. And turned the amplifier on.

I had left the 10ohm, 10watt res's in the rails, and they immediately started smoking when i turned it on. But! It played,a bit quietly, but nevertheless there was music. So i turned the setup off, heard a slight increase in volume for a second (PS cap's discharging?) and then nothing. Took the 10ohms out, and ran it again. It worked, softer than before this time, and in about 3 seconds the one resistor on my PS had started melting. Again, when turning it off, I heard a louder volume. I ran it one more time, with the 3.3ohm res's shortcircuited, and promptly burnt out my 82ohm res on the amp itself, but it played for a second (The res went up like a little supernova.. was cute). And thats where I stopped. Doesn't seem to be any further damage, no heat elsewhere on the board via a little touch observation.

i started troubleshooting. My problems are as follows: My PS puts out +-40V; my transformer it turns out is not the 26-0-26 it has written on its side, but 30-0-30v. Which is bad on the manafacturers part, but I should have tested it. I have a problem there tho; the transformer was sold to me at a discounted "your a student, i pity you" price, but with no paperwork. I'm hoping if I return it, I can get it properly wound, without any conflict or bad business practice on their part. But jumping the gun a little; if I cant, can this amp run at +-40Vdc, is that what caused the burn out?

My next error, my very bad, was that I forgot to link the zobel earth point to ground. I'm not too sure if that is a big problem or nothing to be concerned about?

Other than that, I cant think what else could have caused it. I found it weird that it was the negative rails in both case that burnt out, the positive rails ran as cool as a whistle, what does that point to?
I'll poke around some more til I get a reply, I should have a batch of photos up later tonight. I'll hook up my second amp too, but I'll test it later.
what is the rated mains input voltage?
What is the actual mains input voltage when you measured 30Vac?
Is 30Vac when open circuit?
26Vac is rated voltage when supplying rated AC current into a resistive load.
Regulation is [30/26-1]*100=15%
What is wrong with that?
Hehe... I did test with a light bulb, very quickly tho, just to make sure it lit it, but nothing bad happened in that short time, it lit. Additionally, when the inputs where shorted, and no load on the amp's output nothing burnt out. Only when I added a speaker load did the PS burn out. Maybe I'll try my other PS.

Rated mains voltage in SA is 230Vac, i admit i didnt measure the actual mains voltage, i'll check later.
Yes, it did measure 30Vac with the PS open circuited, and the transformer itself open circuited; it was still running through the PS, but thats not a resistive load? So it should be a 26vac transformer then... Sigh, well I cant blame the transformer anymore :( will have to poke around with my DMM.. still, anyone know why only the negative rail heated up, not the positive rail, but the amp still played music?
Thanks Andrew
For sure you made some kind of error, as the vanilla dx has been assembled by many guys so far...

My guess is some sub-par soldering....seeing that it did play for a bit...

Did you check DC rail voltages before connecting?

I am sitting with a soldering iron a.t.m. so you will have to google how to do a diode test on your transistors... I suspect you fried a few...and with the amount of current that tried to go the the front of the amp, I guess your error was there...

Remove all transistors and do the test on each one...

If you are useing the greg erskine layout, confirm that you did not install the obsolete marked parts... and to which points you connected the bias servo..I thin kthey were marked A Band C.

Do not feel bad, if you stick to this hobby, you will go through this many times still.

Next time listen to andrew and put a lightbulb in series with one of the AC wires before the transformer... if it lights up and stays up, power down and look for the error first...
hmmm... thats quite sad.

I checked the rail voltages afterwards only.. I presumed they would be right; they're +40, -40 wrt ground.

I'm not too sure the transistors have all blown.. there was very little heat, and the amp was playing music until that 82 ohm burnt out. Nevertheless, I'm going to move onto the other channel; its the better built amp anyway. I have to unfortunately slow down from now; I admit I rushed a bit, from this weekend on I won't have much time, so these updates will be far slower.

Didn't install the obsolete parts. When I post those photos, you will be able to see my board; it look pretty standard. I'm still wondering as to why only the negative side blew... I will investigate that further. Just to check; getting 30-0-30Vac out of the transformer without a load is indicative of a 26-0-26Vac rated transformer; my transformer is correct?

Will test the second PS properly first; I'm at a loss, as my first PS still works properly, with +-40V rails with those 3.3ohms shorted, so I dont understand what can be wrong with it..

Will play around with it, update later.
You can not tell if a transistor is broken, simply by looking at it...

The front end normally only draws a few milliamps over this resistor, If it is pulling enough current to the front, to burn it, I suspect something is wrong... It does seem to be an achilles heel of the design... I see...

You realy need to check each transistor... hope you have a desolder pump.

Did you check DC offset before connecting a speaker, and also check for AC on the output for any blatant oscillation...?

Try to underbias it a little next time before going for 80mA... around 56mA is perfect and will suit your heatwsinks better.

The bias servo needs to be checked out as well as it's transistor...
measureing its resistance while turnig the trimmer should give a range of about 1k to 3k if memory serves me right...

This transistor has to be electricaly isolated from the heatsink (just stateing the obvious things we need to check first - don't feel insulted)
hehe, not insulted, thanks for the help. I'm well aware you can't tell if a trans is damaged by inspection.. I just was being hopeful...

I have a desolder pump, been a saving grace many times on many projects. Unfortunately, my soldering iron is quite old, a bit rickety. Still gets the job done, but I'm gonna buy a nicer one next week. Still, checking that board wont be the quickest of tasks; I'm going to attempt to get the other board working first.

Will take into account all that you said. Basically, I was very rash and rushed into starting the amp; think I'll be a whole lot more patient from now on. If its any aid, my second power supply runs a light for extended periods of time; I was manually holding the leads on, will connect them properly tomo and do some voltage under load checks, but it looks much better.

Unfortunately, its late, and I'm lacking sleep. Tomorrow is a long day, hopefully I can get some time to check the transistors tomorrow night. Thanks again for all the help, I'll let you know about developments.
Havent got much done today, but I did some load tests on my second power supply. It lights a globe, not too brightly tho, and the measured rail voltages are still 40-0-40 with the globe load. I dont think its much of a load though, so it's not too suprising.

I intend to starting over again tonight with the next amplifier; I just want to check that the instructions on Greg's site are all I need for biasing and starting this amp.
A very basic checklist of what I'll do

Connect PS to Amp, via 10ohm res's.
Short inputs.
Turn on.
Set bias, following Greg's instructions.
Turn off.
Unconnect the 10ohms.
Untie my inputs.
Put a speaker on the output; one lead to OUT, one to ground.
Input a signal into +-IN.
Turn on.

Anything missing?
Yes.. so this is the deal with the second amp.

Set it up for biasing, started it. PS rails are 39.5V and 39.4V, i think that is expected without load. But the biasing wont work. The values, from startup remained the same, looking somewhat right and somewhat wrong; the negative rail is 100mv below +ve rail, better than the first amp, but the voltages are very very low, 100mv and 200mv respectively, and no matter of adjusting will change it more than +-10mv

As for the DC offset, a bit more success there.. I think. I've adjusted it down to around 6mv, could probably go lower, but it doesn't sit very still.. My DMM maybe? Additionally, It is quite a lot higher on amplifier startup, around 15mv, before it drops down, rather slowly; within bounds of the safe 20mv suggested on the website.. I'm not sure if thats the behavior I should be expecting. Anyway, nothing burnt up this time, so at least I can be happy for that. I'm not going to try anything else until I can get a little advice on these voltages though. So, is there hope, or is this amp gone too?
Hi there.... I am not too sure about your Vbias board... can you just post in the DX thread (the original long one) with your voltages and resitors used for bias servo...
I remember carlos advised alternative values to me when I worked with non standard rails,

Yes the negative rail draws less current by design... that is perfect.

Cheap DMMs and low battery status leads to problems at the ends of the ranges... at least it is not showing any speaker frying voltages... Then again most of the transistors in your amp are still switches off at the low bias you are manageing right now... sure that is just a technicality.

When that is sorted out, we jump to the next tune up.. the gain setting, as I found low power sources like Ipods etc can not drive the amp very hard at the standard setting..

Next time remove the rail resitors before you actualy listen to music.... at those rails, with a 10W resistor you can not realy pass that much current before reaching fire hazzard status...

First get the bias fixed retest offset and AC on output..
The slow dropping of offset is standard.
I've just posted a reply to Carlos in the DXamp thread, but for completeness:

I have probably made an error. A check and recheck of the board tells me I've done nothing wrong in terms of components; even used the gallery on Greg's site to recheck layout, mine's pretty standard.

So I've narrowed my problems down to 2 things; firstly bad soldering. I will be getting a new iron soon, my old one is really bad all things considered, so I'll probe around with a DMM and pretty much resolder everything.

My next problem; possible fake transistors. Nordic, those 2sc5200's that were fake from mantech, were they complete write off's? and more importantly, were they marked toshiba? Because mine are marked that... anyway of testing for genuineness?
sigh... In all my re-reads of this topic, I managed to miss your first post where you mentioned fake toshibas... sigh sigh... always just looked at that picture, never realized that it was a toshiba.

I'm thinking thats where my problem lies, do they really have just one pin..? Are you sure they're fake

So... where to find authentics. And! I'm trying to figure out if I've damaged anything else because of the fakes, at least on the second board... I'm not sure tho?
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.