Int. Rect. Class D reference designs

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Has anybody actually built/heard any of the International Rectifier Class D designs? They are what they call self-oscillating PWMs. What is interesting is that the analog front end is a simple op-amp- cheap. The rest is a few logic gates and a few logic inverters. Again very cheap parts. Add a gate driver, two MOSFETs and the output filter and you have the entire Class D amp. Of course the power supplies are somewhat complicated given the gate drive needs. Another interesting feature is having the output FETS inside the feedback loop - and the loop is digital not analog. Its' like a big high power 1 bit sigma-delta A/D. If the distortion figures are really true I am really impressed. Check it out. See www.irf.com/audio. Follow the thread to the reference designs
BTW: I do not work for IR. Just an interested bystander.

Bill
 
I only found two designs using the hysteretic pre filter-feedback topology, is this what you mean? Have a look at the UcD topology which includes the filter in the feedback loop, using feedback to keep inductor and load impedance deviations under control. You also avoid the risk of output filter resonance which could cause extreme voltages at the output with no load.
 
Actually I did take a look at the UCD180. It appears to be a similiar design except there was no schematic on the webiste to take a real look into what they are doing. Having the feedback after the output filter says to me that it is in the analog domain. I've always thought that some feedback is better than no feedback. Although there are Class D amps that do run open loop. Also, the UCDxxx amps appear to use a gate drive transformer, which may not be the best approach of handling the extremely narrow pulse widths. Luckily narrow pulse widths occur only at maximum positive and negative output, when there is lots of S/N. A gate drive transformer does simplify the power supply issues. Overall, the UCDxxx look really good, especially from a distortion standpoint. Seem pricey, IMHO.
 
The UCDxxx series does not use gate drive transformers, look at the link in my signature, there is a schematic for a UcD I have designed/built there, just ignore all the current measuring electronics in the power stage, it contributes nothing to the principle of work. I also think that http://www.ciaudio.com/ucd_aes.pdf explains it a bit more. I dont think there are any too narrow pulses in the UcD as both on- and off-time varies with output signal level, the higher the output level, the lower the frequency (takes longer time to charge the output inductor to a certain current if you have fewer volts to apply across it)
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.