Hello
Currently working on a layout for an IRS2092 based class D. It uses 3 pairs of output FETs. Each driven by totem pole (I think thats the name) transistors.
I have a three questions:
1) Trying to achieve a compact layout. It occurred to me, that if, the FETs and the IRS2092 are through hole components then there would be no need for vias.Both sides of the PCB could be utilised. Is this a good idea? Would the through hole components add to the inductance of the
2) I want to add snubbers to the output. I am happy with calculating a single snubber. What would happen to the calculations with three snubbers in parallel?
3) An advantage of through hole fets is the ability to mount bus capacitors, very close to the FET, on the bottom layer connected between ground plane and the FET.
If anyone could take a minute to answer these questions, it would be very much appreciated.
Many thanks
P
Currently working on a layout for an IRS2092 based class D. It uses 3 pairs of output FETs. Each driven by totem pole (I think thats the name) transistors.
I have a three questions:
1) Trying to achieve a compact layout. It occurred to me, that if, the FETs and the IRS2092 are through hole components then there would be no need for vias.Both sides of the PCB could be utilised. Is this a good idea? Would the through hole components add to the inductance of the
2) I want to add snubbers to the output. I am happy with calculating a single snubber. What would happen to the calculations with three snubbers in parallel?
3) An advantage of through hole fets is the ability to mount bus capacitors, very close to the FET, on the bottom layer connected between ground plane and the FET.
If anyone could take a minute to answer these questions, it would be very much appreciated.
Many thanks
P
I have an all though hole design and it works fine.
I think you can get decoupling almost as close or as close as makes no difference as you can with surface mount components.
I think you can get decoupling almost as close or as close as makes no difference as you can with surface mount components.
It's commonplace to use thru-hole and SMT components in the same design, for whatever reason. Some components are only available in one or the other.
Any conductor has inductance, even a via, but in most cases (for vias) it's trivial. Google 'inductance of a via'.
I don't really understand what you're getting at here, so I'll pass on that one...
True. In general it's possible to achieve greater freedom from strays with SMDs than with thru-hole, particularly with chip capacitors vs. leaded. It's more significant the higher the frequency, of course. A switching amplifier is not a microwave radio or even a PC motherboard.
Thank you for the responses.
Regarding the snubbers, I am concerned with smoking the capacitors. I have noticed that the resistors in snubbers are quite high in their power rating. Would suggest high ripple currents? This why I thought about using three in parallel to share the load
Will google via inductance and have a read. For some reason I've become very anti via. Spend ages moving componets around to avoid them.
Been forced down the smd route (used to be hardcore through hole) by the fact that a lot of newer devices are smd only. For example, the TAS5631, which i plan to use for mids and highs. They don't have the desired characteristics for sub amps. Can't see myself having a sub woofer with a 2 ohm impeadance. Hence, the IRS2092.
So, using through hole components isn't a significant disadvantage at the kind of frquencies the 2092 runs at. Is this a fair assessment?
This is my first venture into class D. Want to make a class D equivalent to my class AB setup and see if I can tell the difference.
Regarding the snubbers, I am concerned with smoking the capacitors. I have noticed that the resistors in snubbers are quite high in their power rating. Would suggest high ripple currents? This why I thought about using three in parallel to share the load
Will google via inductance and have a read. For some reason I've become very anti via. Spend ages moving componets around to avoid them.
Been forced down the smd route (used to be hardcore through hole) by the fact that a lot of newer devices are smd only. For example, the TAS5631, which i plan to use for mids and highs. They don't have the desired characteristics for sub amps. Can't see myself having a sub woofer with a 2 ohm impeadance. Hence, the IRS2092.
So, using through hole components isn't a significant disadvantage at the kind of frquencies the 2092 runs at. Is this a fair assessment?
This is my first venture into class D. Want to make a class D equivalent to my class AB setup and see if I can tell the difference.
The problem with any kind of switcher is the high DI/Dt switching...
Some interesting links on this thread
http://www.diyaudio.com/forums/power-supplies/193705-switcher-emc-design.html
Some interesting links on this thread
http://www.diyaudio.com/forums/power-supplies/193705-switcher-emc-design.html
Sometimes it's not the operating frequency that is the issue but the way a circuit responds to stray RLC and physical layout the causes problems in analog audio designs.
Mike
Mike
- Status
- Not open for further replies.