P68 Overheating!!

[led_zeppelin]

Hi again... i've just been breadboarding a p68 and it seems to work ok, but the first set of outputs get really hot... (q7 and q8)

I'm using +- 70v rails with a 625w transfo... i put a 40w bulb in series with the transfo and the rails fall to +-25!! there isn't much dc ofset (about .2 of a v) but the bulb is always lit when i power it up.

It sounds fine with a signal, until you turn it up, when it distorts. i think i've used a few different values for resistors, such as 330 for r13 and r14, and 680 for r5,6 and 7... would it be these? the other outputs stay cool... just q7 and q8 get hot...

Any help will be much appreciated as i need to get it working because it is my a level project!!

Thanks again, as usual.

Matt

[tool49]

I'm not entirely sure, but I believe that this amplifier should not drain 40W at idle. So your light bulb lighting up means that something must be faulty in your breadboard circuit. But then again, I'm no expert!

I would revise the whole circuit again. And are your driver transistors on a common heatsink with the diodes? If not, it could be thermal imbalance.

Hope this helps!
Sébastien

[hacknet]

i`ve not seen the schematics but is there a bias pot where you can lower the idle current?

[led_zeppelin]

no... it uses a bootstrap principle with two diodes...

I have been looking at the schematic, (Project 68 ESP ) and i don't quite get it...

basically, the base of both q7 and q8 are connected to the rails by a 220 ohm resistor.... now i measure around 40v across the two bases, (obviously with severely sagging rails) which is, i'm guessing where my problem is.... should this be like this?

(i don't really trust my volt meter as it is a cheap *** digital one, but i measure about 0.5 on the bases of q5/6 which i presume would be right, so why the +- 20 odd volts on the base of q7/8? i can see that they are simply attached to the rails.... i cant understand... ****Newbie Alert****)

[AndrewT]

Hi Led,
The esp site warned not to try this project as a beginner.
Q7&8 are part of a 3 deep sziklai output stage.
the correct voltage across the bases should be +Rail to -Rail voltage less 2 Vbe. If your supply rails have collapsed to +-21v then the base voltage wiil be as measured ie. 40v.
Post 2 said breadboard again, I agree.
BTW a number of similar threads and authors have pointed out that this circuit is more prone to oscillation. Less than tidy breadboarding with long wires may exagerate this problem. To relieive this I believe ESP warns to use the PCB designed for this project.
regards Andrew T.

[hacknet]

please correct me if i`m wrong, i`m another noob.

i did a couple calculations at it works out that the amp might be working correctly...

i suppose you are working on 230v mains and thus the 40watters would have a series resistance of around 1.3kohms.

with 1.3kohms in series per rail, a small current would cause the bulbs to light because of bulbs high resistance and the amp`s constant currents.

you experienced a drop of 45v per rail. this would roughly put bias at around 35 milliampere. 35ma Iq is about right with saggy rails. my guess is it would rise to ~65ma at full voltage, giving you a idle dissipation of around 10 watts... (i might be wrong here, can elliot confirm? )

try replacing the 40watters with a 10ohm 10w resistor. it give enough time to hunt for flaws without nuking the output transistors.

[led_zeppelin]

does that mean that the trannies are meant to cook? and i'm sure something is wrong, because before using the big transfo, i used a small 19-0-19, and exactly the same happened... with no current limiting bulb, it turned into a 1 bar electric fire on the two q7 and 8, so i bolted them to the heatsink, to stop them dying and it just got hot. they shouldn't be switched hard on should they?

can someone explain why they are connected straight to the rails then? does the stage before turn them off then?

I actually get a good sound out... untill you turn it up... i don't dare unleash the 625w as i fear the heating transistors would turn to dust.

[led_zeppelin]

also, i would love to use esp's pcb, however firstly i cant afford it, and secondly, i need all my outputs in a line as that is how i need them mounted on my heatsink.

Not to mention i can make the pcb's in 6th form for nothing... so i just want to check it will definatley work. No one will let me see a copy of esp's, so i can verify my design, which is fair enough......

My breadboarding is very neat, i've sticked to amp rules, using one ground point etc.... and i've built the amp a few times... i just thought it may be easier to look at the schematic and find out what would cause them to melt, rather than poking my fingers into 140v dc..... which may not be sensible could i not just use a more standard output design, like simply omit q7/8, and basically use the push pull driver (with maybe 10 ohm emitter resistors) with the base of the outputs connected where they are? what exactly do q7/8 actually do?

[led_zeppelin]

sorry hacknet... i think you may have mis understood... the 40w bulb is in series with the primary of the transfo, not the secondary.... so it comes on as the capacitors charge (10,000 uf per rail) and then should go off, but it is sitting on... not completely at full, but still more than it should be... and this is obviously only going into my heatsink, as nothing else gets hot.

[Rarkov]

Urm, a 3dB increase in is double the power. Four times the power is 6dB which is double the perceived volume.

I could be wrong - but I think you're way off with your 10x!

Gaz