TA7317P output protection

with

There

I was wrestling with a similar function C1237HA / uPC1237HA recently, as falsely holding off the relays, replaced associated caps, replaced the device , still the same

Good Morning,

I've usually seen these IC's in NAD receivers. I've certainly had issues with both the capacitors and resistors going bad, even though they seem to measure OK in circuit. I'm going from my memory (bad idea) but 150K, 180K, 220K 15K, 18K and 22K stick in my mind.

Best of luck, Tim Schwartz Bristol Electronics

with

There

For a Panasonic SA313 use of TA7317P, p1 to p9

-.7,0,0,0,-.7,1.3,0,1.3,3.1 V dc

On Fri, 24 Feb 2012 23:43:18 -0500, "John Smith" put finger to keyboard and composed:

It would have been nice for you to have made it easy for us ...

AISI, the maximum voltage at this point would be equivalent to the voltage drop across 4 forward biased diodes. Simlarly, the voltage drop at pin 5 should be equal to a single forward biased diode.

If the voltage drop is taken to be 0.8V, then this is consistent with both measurements.

What voltage do you see at pin 5 of the replacement? Do pins 1,2,3 and

7 of the replacement part now show the expected voltages?

Does pin 9's series resistor (R372) measure correctly (22K)?

- Franc Zabkar

--
Please remove one 'i' from my address when replying by email.

On Fri, 24 Feb 2012 23:43:18 -0500, "John Smith" put finger to keyboard and composed:

AISI, the voltage across R372 (22K) must be 73.3V.

Therefore its power dissipation would be ...

73.3 ^ 2 / 22K = 244mW

However, the parts list rates it for 1/5W.

- Franc Zabkar

--
Please remove one 'i' from my address when replying by email.

Thanks but this is one of the first things I checked before replacing the device. R372 does measure 22K. I actually cut one of its legs to test it out of circuit because it's a pain to get the board off the heatsink (all output FETs have to be unsoldered) I then soldered its leg together again.

I spotted another potential design error. Compare the TA7317 Vcc max with what I measured. This only affects pin 6.

I'll find out more when I get back to it on Monday.

old guy

Sorry, I didn't think a quick cut & paste into Google was hard.

I'll be able to answer your other questions on Monday.

old guy

...

...

Useful information. Thanks.

At a glance, 0.6 volts at pin 8 is wrong - capacitor C312 should charge up through R364.

If it is held low, the thing will never turn on. The voltage at pin 9 looks like may be low - R372 could be way high in value.

Mark Z.

On Sat, 25 Feb 2012 19:00:43 -0500, "John Smith" put finger to keyboard and composed:

No harder than a quick cut & paste into a newsreader ...

BTW, I notice that the voltage rating of the filter capacitors (C401 / C402) on the amp's positive and negative supplies is 80V. Are you sure the AC mains voltage selector is in the correct position?

- Franc Zabkar

--
Please remove one 'i' from my address when replying by email.

....

I'll check R364/C312 tomorrow. R362 measured 22K out of circuit. I might replace it anyway.

I'll check it on Monday. I'm in a 115V 60Hz area.

On Sun, 26 Feb 2012 08:11:41 -0600, "Mark Zacharias" put finger to keyboard and composed:

AISI, pin #8 should go low if either pin #1 or pin #7 are high, or if either pin #2 or pin #3 are outside the range of -0.7V to +0.7V.

If there is a fault condition, one or more of Q6, Q18, and Q19 will be on. This means that Q10 will be off and Q11 will be on. This in turn means that Q12 and Q13 will be off and the relay will be open.

- Franc Zabkar

--
Please remove one 'i' from my address when replying by email.

It took me a little longer than expected to get back to this but here's the story so far.

The mains voltage selector is set correctly for the area I'm working in. So

+75V and -75V rails seem to be correct.

I removed the replacement TA7317P and fitted socket strip.

With no chip fitted R372 = 21.6K and R364 = 220K.

All other resistance checks made sense and no adjacent socket pins are shoreted together.

I then connected pin 6 to pin 4 of the socket to turn the relay on and also connected four 1N4148 dioded in series from pin 9 to pin 4 of the socket.

With the power on, here are the voltage readings I now get. Note that no chip is fitted, just the diodes mentioned above.

1. -15V
2. 8.8mV
3. 0.17V
4. 0V
5. -75V
6. 0V
7. 0V
8. 2.67V
9. 2.7V

The voltage on C314 is varying between -42 and -49V, presumably due to ripple. I'll scope it tomorrow. The voltage at pin 1 of the socket is also varying a bit too.

Thanks for any further suggestions.

My next move will be to test the remaining TA7317P I bought, with a 22K resistor to pin 9 and pin 4 grounded with no other pins connected and see what I get on pin 9.

Old guy

On Thu, 1 Mar 2012 14:30:32 -0500, "John Smith" put finger to keyboard and composed:

That means that quite a few components are operating close to spec or slightly in excess of spec.

I'd change the capacitor, but I don't believe that level of ripple would be enough to influence the circuit.

AISI, pin #1 senses the current in the 0.33 ohm load sharing/sensing resistors R33, R34, R35, and R36. It would require about 2A to turn on Q324. Q325 would then turn on after a delay determined by C311 and R357, sending pin #1 of the TA7317P high and causing the relay to open.

I suspect that your replacement parts may have a string of 3 diodes on pin #9. If the voltage at pin #5 is -0.6V, then this would tend to confirm it.

I notice that the datasheet identifies each diode as a Q rather than a D, so I suspect that the original part uses diode-connected transistors rather than actual diodes. Perhaps your replacement part uses the latter?

- Franc Zabkar

--
Please remove one 'i' from my address when replying by email.

On Fri, 24 Feb 2012 23:43:18 -0500, "John Smith" put finger to keyboard and composed:

I believe pin #3 is being used to provide thermal overload protection. That said, I can't see how it could possibly work, even with the "correct" supply voltage at pin #9.

Here is the relevant section of the circuit:

pin #3 o | pin #9 o-|-- R366, 15K -- TH301 --|-- R367, 3K --o pin #4, Ground | | |-- R365, 180K ----------|

TH301 is an NTC thermistor, Panasonic p/n ERT-D2FHL503S:

It's resistance at 25C is 50K.

AISI, in order for the thermal overload to activate, pin #3 would have to exceed 0.6V (the B-E voltage of Q15 in the datasheet). However, even if we allow the resistance of TH301 to fall to 0, then the voltage at pin #3 is still only 0.57V, assuming a voltage of 3.2V at pin #9. At 1.8V, the voltage at pin #3 would be only 0.32V.

C316, R371 and R374 appear to provide a dethump delay at switch-on.

- Franc Zabkar

--
Please remove one 'i' from my address when replying by email.

....

Thanks Franc

I haven't checked your analysis but in my time as a circuit designer (not saying how many years or what I worked on) I encountered many situations where pleasing management was the most important thing.

So a thermal overload test may have been done on one unit and it may well have passed. And then it may have been assumed that...

The unit has a very big heat sink, actually I don't recall many that were bigger. So I'm not too concerned if thermal overload doesn't work. No one is going to be injured by a FET going pop, well except the speaker if the relay doesn't turn off, so my risk assessment is that it doesn't matter as long as I can get the relay to turn on when the output isn't stuck at a rail.

If I can't do that then I won't get paid for the repair :(

Old guy

On Fri, 2 Mar 2012 00:52:36 -0500, "John Smith" put finger to keyboard and composed:

I pointed it out because it looked wrong. In fact the design seems to have too many oddities.

That said, the 4 protection inputs look OK to me, so there does appear to be a problem with the IC.

There is one thing you might like to try. The TA7317P datasheet specifies a maximum current of 5mA into pin #9. The circuit as it stands draws about 3mA through R372. Perhaps you could add a 47K resistor in parallel with R372. If the voltage on pin #9 rises, then this would point to an internal regulation problem.

- Franc Zabkar

--
Please remove one 'i' from my address when replying by email.

Thanks Franc,

Here are the results of todays experiments.

I took another of the TA7317P I bought and carefully bent most pins so that when I put it in the socket only pins 4,5 and 9 are in the socket.

Here's what I get:

Socket pin V, Chip pin V.

1 -16.5, -1.04 2 0.01, -1.98 3 0.01, -0.66 4 0, 0 5 -2.39 -2.39 6 76.6 -2.38 7 0, -1.29 8 0.3 0.122 9 0.3 0.3

Something tells me that this can't be a TA7317P despite its markings.

Maybe the manufacturer got it reversed in the package. I think it should be safe to try one backwards.

Old guy