Yet another version of IR audio amp ??

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi, to all people who are there, and willing to help me figure out some issues with this amp.
I've built pcb for 2 channel with schematic revised by pat allen and I added oc protection and those parts were altered:

MURS120 is in mine desing SBYV27-200 15ns/200V
irfbn23n15d -> irf640
ir2010 -> ir2110
all mm transistors are 2n
as IRF suggested heatsink is electrically conected to ground.

I did tested mine amp at +-25V and the main issue is that it drawns about 0.34A at +25V and 0.44A at -25V and inductors are cold, but mosfets are getting worm after 20sec not too much but . . . and frequency is 400kHz but if I try to set some lower then current drawning also drops a little bit. (0.2A and 0.3A)

So I would like to hear some oppinons about this and btw I am hopeing that is not issue in mine pcb layout . . . . .
 

Attachments

  • dsc02023 (700 x 525).jpg
    dsc02023 (700 x 525).jpg
    39.3 KB · Views: 2,731
What do you want to tell us with your last statement ? That you would NEED a layout with all grounds connected through to a groundplane or that you HAVE one (though it seems that the latter is the case - which is very good) ?

As far as the output signal goes: I have seen worse on class-d prototypes (but it could be a little better) ! Have you already played with deadtime and/or snubbers ?

What I definitely don't like are the glitches on the IR2110's inputs. Such things can lead to increased losses and/or driver failure.

You would now have to find out how they get there. A few possibilities :
- capacitive coupling
- ground bounce (is your groundplane interrupted somewhere ?)
- inductive coupling
- bad PSU decoupling (what do the PSU rails look like on the scope?)

Congrats on an otherwise very nice looking device.

Regards

Charles
 
Member
Joined 2005
Paid Member
Zkaiser,

I had time to go over your schematic today, so here are a few comments:

- gate resistors to the output MOSFETs are too low, in all of my simulations using the IR driver ICs I use 47 ~ 100-ohm, 10-ohm will result in very high cross-conduction currents.
- IR2110 needs to be fed with 15-volt instead of 12 as it'll give you just a little bit more headroom and better on/off characteristics for the MOSFETs.

Best regards,

Sander Sassen
http://www.hardwareanalysis.com
 
phase_accurate
I have ground plane without breaks, but pcb layout which I sent to you was without gnd-planes because I wanted to everybody be able to see all wires on mine board in one picture, so I DO HAVE groundplane on both sides of my board.

btw I didn't played yet, with deadtime and/or snubbers
(but I should)

When amp starts to oscillate the PSU rails looks very dizzy on the scope, so this is the reason why outputs from 7404 (or inputs of ir2110) look how they look.
If I am not wrong there are 1Vpp noise on +-5V supply for opamp and triangle don't looks too good also.

SSassen
Actually I do use 15V for ir driver and you might be very very right about this resistors because I do have high cross-conduction. Are you suggesting to put higher resistors there just to make deadtime a little bit longer ?

Is there migh be a case that cross conduction cause excessive noise production on 5V and 15V rails ?
 

Attachments

  • dsc02042 (1000 x 750).jpg
    dsc02042 (1000 x 750).jpg
    72.9 KB · Views: 1,691
Member
Joined 2005
Paid Member
Zkaiser,

Is there migh be a case that cross conduction cause excessive noise production on 5V and 15V rails ?

That's a clear sign of cross-conduction, change the MOSFET gate resistors, start with 100-ohm and gradually reduce them untill the MOSFET's are relatively cold and you have the least switching noise on the scope.

Best regards,

Sander Sassen
http://www.hardwareanalysis.com
 
I have tried with 47ohm resistors and now the current is for about 0.1A lower than before, 0.2A for +25 and 0.3A for -25V but I still have glitchfull signal on input of 7404 and I thik that reason for this is in commont reference points, so I need to make another route for -25V or logical ground for 7404 and ir2110 to resolve this supply fluctuations.:eek:
 

Attachments

  • pcb layout (1000 x 565).jpg
    pcb layout (1000 x 565).jpg
    83.1 KB · Views: 1,650
Any suggestions ??
I also need to say that this amp sounds pretty good with rail supply +-30V, but I needed to set lower ferq. (200kHz) in order to get less dissipation on ir2110 and mosfets but this is still not good (idle current is 0.1A for +rail and 0.2A for -rail). Of course this is all wrong cos then I have 6Vpp 200kHz noise on output. So there must be something on pcb that mess up +-5V and +5V supplies at freq. of 400kHz ?
 
Hello!

Just to let you know that you cannot replace directly an ir2010 buy a 2110. Logic level of the 2110 is for cmos driving, ie, Vin low is 6V and Vin hi is 9V, so a TTL driver will feed only ir2110 close to their input limit. Try it with a CMOS drivers IC like the 4001 or 4011. Add too some 0.1uf capacitor between +/- power supply of the op amp and the speaker ground, this will cancel noise on your triangle wave. Like another guy told too, rise the value of the gate resistor to at least 15 ohms and drive the ir2110 close to 15V. It should work! Just a pick of my news prototype of a 3Kw 2 ohms amp with ir2110 and APT killer mosfet! Look at theire web site, their mosfet are incredible and best, they work!

Have a nice days!

Fredos
 

Attachments

  • 3kw amp.jpg
    3kw amp.jpg
    54.8 KB · Views: 1,574
hi,

fredos, this make sense but in both datahseet 2110/2010 IR says that they are CMOS/ LSTTL and 3.3V logic compatible

Features
• Floating channel designed for bootstrap operation
Fully operational to 200V
Tolerant to negative transient voltage, dV/dt immune
• Gate drive supply range from 10 to 20V
• Undervoltage lockout for both channels
• 3.3V logic compatible
Separate logic supply range from 3.3V to 20V
Logic and power ground ±5V offset
• CMOS Schmitt-triggered inputs with pull-down
• Shut down input turns off both channels
• Matched propagation delay for both channels
• Outputs in phase with inputs
• Also available LEAD-FREE

So now what I am suppose to think now ??
I did rised resistors to 100ohm and put some 0.1uf caps in critical places and now amp can work pretty clean @400kHz but I still have slight warming on ir2110 and fets.
Are you saying that I need to make some sort of level shifter (adapter) for transition from triple inverter (TTL logic) to cmos levels 1(100%-90% of 15V) and 0 (0% to 10% of 15V) ?

BTW good looking amp do you have some hi res pics of it?

Regards
Zeljko Kaiser
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.