I'm hoping someone maybe able to help. I've got a tamp ta2500 here that has only got two fets switching the rails and are marked t410 and some other number that I can't see at the moment. Every output transistor and one driver as gone short along with one of these fets. There is a wider footprint for a larger device and also a spare space for another two in parallel. I was wondering what the correct replacement for the fets would be. I was thinking of usinig irz44 but maybe it's a good idea to use the footprints for the larger parts?
The funny thing is the fet that had gone short had never been soldered in and must of been making good contact for awhile (friction fit lol) and maybe the main reason for this amp self destruction.
Also I was wondering if you can test class h/g amps safely as a straight a/b amp by just removing the switching fets and just running off the lower rails at reduced power?
Thanks in advance for any help
The funny thing is the fet that had gone short had never been soldered in and must of been making good contact for awhile (friction fit lol) and maybe the main reason for this amp self destruction.
Also I was wondering if you can test class h/g amps safely as a straight a/b amp by just removing the switching fets and just running off the lower rails at reduced power?
Thanks in advance for any help
Last edited:
Hi Kanks, just a couple of observations:
the metal tab of the FET is its Drain (centre pin) so this is often used in preference to the centre pin from a PCB design point of view.
Near the FET's, there should be 1 or 2 FRED (fast recovery diodes) which may be a 3-pin TO-220 device depending on the manufacturers choice - so check you're not confusing the functions of the poorly marked components.
The amp will actually run if the fet has shorted (it'll just run hotter with marginally more distortion).
The Switching fet's i usually see in this type of amp are an IRF640 (which is nearly discontinued).. a TO-220 package is fine for the FET because its acting as a switch - therefore is dissipates only a small ammount of power as its either OFF (no power) or fully ON.
The important things when looking for a substitute are:
Rds(ON) should be low ~0.2 Ohms
Vds should be around 600V as it has to survive the switching spikes
Tr and Tf should be very fast - for the IRF640 they are 35ns.
This is tied in with the gate capacitance (Ciss) which should be around 2000pf.
If in doubt, replace the FREDs too as these are supposed to switch at 35-65ns and you can't really test that without a large hadron collider.
Personally, I run this type of amp up as you suggest…
1. replace the Fet with a Link and run the amp up at a lower voltage.
if it seems good, put the Fets Back in and Run the amp up gradually with scope CH1 showing the audio out and CH2 showing the transistor's power rail.
you should be able to see the rail switching.
avoid sticking your scope onto the FET gates as it will screw up the timing.. if you need the check the gate drivers, do your probing before the gate drive transistor.
hope this is useful info
jonathan
the metal tab of the FET is its Drain (centre pin) so this is often used in preference to the centre pin from a PCB design point of view.
Near the FET's, there should be 1 or 2 FRED (fast recovery diodes) which may be a 3-pin TO-220 device depending on the manufacturers choice - so check you're not confusing the functions of the poorly marked components.
The amp will actually run if the fet has shorted (it'll just run hotter with marginally more distortion).
The Switching fet's i usually see in this type of amp are an IRF640 (which is nearly discontinued).. a TO-220 package is fine for the FET because its acting as a switch - therefore is dissipates only a small ammount of power as its either OFF (no power) or fully ON.
The important things when looking for a substitute are:
Rds(ON) should be low ~0.2 Ohms
Vds should be around 600V as it has to survive the switching spikes
Tr and Tf should be very fast - for the IRF640 they are 35ns.
This is tied in with the gate capacitance (Ciss) which should be around 2000pf.
If in doubt, replace the FREDs too as these are supposed to switch at 35-65ns and you can't really test that without a large hadron collider.
Personally, I run this type of amp up as you suggest…
1. replace the Fet with a Link and run the amp up at a lower voltage.
if it seems good, put the Fets Back in and Run the amp up gradually with scope CH1 showing the audio out and CH2 showing the transistor's power rail.
you should be able to see the rail switching.
avoid sticking your scope onto the FET gates as it will screw up the timing.. if you need the check the gate drivers, do your probing before the gate drive transistor.
hope this is useful info
jonathan
ideally you should use the same device that was there in the first place - are you able to identify it by looking at the other channel ? i've been comparing the data sheets and the IRF640 has pretty much the same Rds(ON) resistance, better voltage handling and better Trise/fall speeds, but it has a lower continuous ON current - which suggests to me that the PCB was probably designed for a pair of 640's, which have been superceded by the Z44.
However, just cant simply swap in a pair of devices because this would double the gate capacitance - requiring the gate driver to be modded to suit this.
from experience, i've found its better to leave designs alone. (unless, of course you want to call it a reseach project ;-)
j
However, just cant simply swap in a pair of devices because this would double the gate capacitance - requiring the gate driver to be modded to suit this.
from experience, i've found its better to leave designs alone. (unless, of course you want to call it a reseach project ;-)
j
I agree wholeheartedly with the advice to stick with the devices that were originally fitted wherever possible. I know it gets difficult when parts are discontinued, and the 'replacements' aren't always a good match on one or more parameters. At least the saturation voltage seems to get lower on all the new parts, but all that does is reduce the power lost in the switch by a few percent, which may not make a noticeable difference to anything important.
A lot of these big amps share a common PCB betwen models, but a lot of the small component values change at the same time, so don't just put 'extra' devices into the boards.
The power devices are cheap, so I always replace everything when I get a faulty unit.
Watch out for non-genuine devices, as they WILL fail.
I have a pair of blown fuses with power resistors mounted across them which I use when powering up a repair for the first time - that has saved me a lot of grief when there's a problem with the biasing - often the cause of the original fault, but not easy to diagnose without the output devices.
I find that the 'protection' circuitry isn't very effective and all the usual problems with shorted speaker cables and overheated voice-coils will result in a major blow-up just like any other amps, so trust nothing.
Graham.
A lot of these big amps share a common PCB betwen models, but a lot of the small component values change at the same time, so don't just put 'extra' devices into the boards.
The power devices are cheap, so I always replace everything when I get a faulty unit.
Watch out for non-genuine devices, as they WILL fail.
I have a pair of blown fuses with power resistors mounted across them which I use when powering up a repair for the first time - that has saved me a lot of grief when there's a problem with the biasing - often the cause of the original fault, but not easy to diagnose without the output devices.
I find that the 'protection' circuitry isn't very effective and all the usual problems with shorted speaker cables and overheated voice-coils will result in a major blow-up just like any other amps, so trust nothing.
Graham.
The correct fets are either:
2 x IRF540 in parallel
2 x IRF640 in parallel
1 x IRFP250N with legs bent to fit.
there may be others if the factory were short at the time but I haven't seen any.
C-Mark fit either of these, I suppose whatever's cheaper at the time.
I wouldn't run these for one second with the fets shorted, they quite happily go bang at the best of times. If you want to try the amp at the lower voltage, pull the HV transformer connectors off the power supply board.
These amps are also not happy running full tilt without a load, the commutating snubber capacitors have a habit of going short.
If I were you I'd save your money and buy something better
2 x IRF540 in parallel
2 x IRF640 in parallel
1 x IRFP250N with legs bent to fit.
there may be others if the factory were short at the time but I haven't seen any.
C-Mark fit either of these, I suppose whatever's cheaper at the time.
I wouldn't run these for one second with the fets shorted, they quite happily go bang at the best of times. If you want to try the amp at the lower voltage, pull the HV transformer connectors off the power supply board.
These amps are also not happy running full tilt without a load, the commutating snubber capacitors have a habit of going short.
If I were you I'd save your money and buy something better
Latest e-mail address
Hi all,
I'm sorry, but I think the change in my email address somehow failed to get through the system.
I ddn't realise till aol stopped forwarding mails from their site to former customers who were no longer with them.
I was still getting newsletters, etc., but not replies to my postings.
Hopefully this will appear at the top of the thread and put anyone who want me on the ruight course.
Graham
grahamprie@sky.com
Hi all,
I'm sorry, but I think the change in my email address somehow failed to get through the system.
I ddn't realise till aol stopped forwarding mails from their site to former customers who were no longer with them.
I was still getting newsletters, etc., but not replies to my postings.
Hopefully this will appear at the top of the thread and put anyone who want me on the ruight course.
Graham
grahamprie@sky.com
I would setup your amp to work without output transistors.
Feedback the output to the inverting input so you get a working loop.
You can then work on the pcb by twisting it gently until the amp goes wrong.
I once had an amp that didn't work right. I twisted the pcb and it stopped working.
I traced it down to a unsoldered joint that had been touching its via intermittently.
Even just twisting the pcb might open up the hair line crack enough to see it ?
Feedback the output to the inverting input so you get a working loop.
You can then work on the pcb by twisting it gently until the amp goes wrong.
I once had an amp that didn't work right. I twisted the pcb and it stopped working.
I traced it down to a unsoldered joint that had been touching its via intermittently.
Even just twisting the pcb might open up the hair line crack enough to see it ?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.