Devil in the detail ... missed to update snubbers in the BOM
The devil's in the detail. Really!
With the changed more heavy MosFets for higher voltages also the snubbers needed an adjustment, but I missed that in the BOM!
Take care, not just the values changed, but also voltage and power ratings.
Here the updated BOMs for 2x65V and 2x80V.
Settling an error free BOM often needs some loops, in case you notice/suspect errors, please let me know.
The devil's in the detail. Really!
With the changed more heavy MosFets for higher voltages also the snubbers needed an adjustment, but I missed that in the BOM!
Take care, not just the values changed, but also voltage and power ratings.
Here the updated BOMs for 2x65V and 2x80V.
Settling an error free BOM often needs some loops, in case you notice/suspect errors, please let me know.
Many thanks goes to Markus for sharing this ultra nice amplifier with us.
Have a look at posts #
- 242: bridging adaptor with gain
- 364 and 365: bridging adapter without gain
BR, Toni
Last edited:
You mean building 2x65V_IRFI4020 and simply replace the IRFI4020 with the IRFI4019?
Change of MosFets is always good for surprises, but from theory both devices appear close enough that it should work and I think up to 2x60V it should be fine even with a 4R load and unchanged overcurrent setting.
Dear Markus,
this is exactly what I meant. I have "tons" of unused IRFI4019 from an unfinished IRS2092 project and this would be a very good option to recycle them.
Br, Toni
Very nice build Chocoholic,
Can someone please point me towards the post showing Amplitude Vs Frequency response with variety of loads say 2, 4, 8 ohm and also open load response. I am curious to see the behavior of this post filter feedback. I have seen the pre-filter feedback response gives peaking at HF.
Can someone please point me towards the post showing Amplitude Vs Frequency response with variety of loads say 2, 4, 8 ohm and also open load response. I am curious to see the behavior of this post filter feedback. I have seen the pre-filter feedback response gives peaking at HF.
Nice one Marcus!.
Now if only someone could design a "safe" cheap to construct SMPS to pair with this project.
If our Rand currency wasn't so weak 1R = $14,I would be getting involved in another build, but economics is hutting our wallets at the moment in South Africa
Nice one Marcus!.
Now if only someone could design a "safe" cheap to construct SMPS to pair with this project.
If our Rand currency wasn't so weak 1R = $14,I would be getting involved in another build, but economics is hutting our wallets at the moment in South Africa
I could offer two unregulated SMPS units at 120US$ for both units (exclude shipping), 600W Continous,, 1000W Peak.
Size is 12x11CM
Regards
The filters in front of my QA400 are to slow for a correct measurement.
From design theory the 2x40V version shows a flat gain of 13.9db (+0.1db / -0.5db) up to 26kHz.
-3db bandwidth is larger than 60kHz with any load.
The graph is showing gain and phase at 2R, 4R, 8R, 99999R.
Attached also the LTspice file with the averaged model for AC analysis.
Feel free to simulate the other versions.
Attachments
Very nice Choco,
looks like Zobel is mandatory part of the response, omiting it causes peaky response with open loads.
Very nice system transfer function...excellent work! And thank you for continuing to enlighten us by discussing the specifics of your design methodology. I purchased a set of PCBs but I have promised myself that I will not begin assembly until I have learned the details of how this design is functioning. Believe me, it's a little strange to go back to my university text books from 25 years ago to relearn things! I have been trying to simulate the feedback loop function for the LiteAmp using the bode plot example that you posted back in the SystemD_2kW thread, but I am having issues. Using the averaged model I am seeing the same result as the model I had constructed...fs=175kHz? I must be missing something...
Yes, especially for the 40V version I squeezed for high loop gain up to high frequencies. Besides reasons from control theory I also used the Zobel for carrier shaping in order to achieve lowest THD.
Using the 2kW model and translate values in order to get the behavior of the LiteAmp really is a difficult stretch.
In general please note:
The averaged model is not suited to simulate the switching frequency.
In fact the averaged models with the original values, which I have posted do not oscillate. You can use it for transient analysis and will get a correct reproduction of the signal without oscillation.
It will only oscillate if you change values and violate the Nyquist criteria.
In order to look into fs and related topics you can use the model in posting #289:
http://www.diyaudio.com/forums/class-d/255046-systemd-liteamp-29.html#post4342759
Wow - really like all the work shared here. Thanks ChocoHolic.
I would like to build one of these amps, with 8 channels for home theater- combined on 4 boards with the gain board integrated.
After reviewing the schematics I had a couple questions, I was hoping someone could answer:
1) What should the sync line be tied to? Is this for multiple channels?
2) The gainboard is differential input. If a single ended input is desired would you suggest tying one side to ground, adding a balanced to unbalanced converter, or redesigning for single ended? It seems like optimal solution would be to have an input that could be either - would that work?
3) How difficult was the layout to get right? App Note 1135 from IR goes into a lot of detail that hints at significant trickery -
Thanks in advance,
Scott
I would like to build one of these amps, with 8 channels for home theater- combined on 4 boards with the gain board integrated.
After reviewing the schematics I had a couple questions, I was hoping someone could answer:
1) What should the sync line be tied to? Is this for multiple channels?
2) The gainboard is differential input. If a single ended input is desired would you suggest tying one side to ground, adding a balanced to unbalanced converter, or redesigning for single ended? It seems like optimal solution would be to have an input that could be either - would that work?
3) How difficult was the layout to get right? App Note 1135 from IR goes into a lot of detail that hints at significant trickery -
Thanks in advance,
Scott
The sync input is an optional feature, if you you do not use it-then you
can tie it to GND (leaving open also possible, but tied to GND it will pick less disturbances).
If you want to use it then i.e. you could use a master clock which
feeds the lite amp with 384kHz, the DAC with 192kHz and the PSU with 96kHz.
Third option is to connect all sync input of lite amps together, which will lead
to a natural sync of all chanels. It will depend on your detail implementation whether this is more fortunate or not compared to tying them GND.
Yup, that's perfectly Ok. You can use the inputs as
single ended inverting (noninv tied to GND), single ended noninverting (inv tied to GND), or differential inverting, or differential noninverting.
All modes are perfectly fine and of course with the above options you can also configure bridged mode. No need for another front end.
Of course you can skip the complicated limiter, if you do not need it.
Layout makes 30% of the R&D knowhow for classD.
It's all about 3-dimensional field theory, loop inductances and inductive couplings, track capacitances and capacitive couplings... last but not least also about antenna theory.
If you have this knowhow, then difficult is not the right word, it's more like a complex puzzle and some experience about fortunate trade offs.
If you do not have this knowhow, then it will take you pretty some time and design loops to learn. Opinions about the 'right way' diverge - IMHO there are multiple reasonable concepts, each with different advantages and disadvantages. ...and unlimited amounts of unreasonable approaches...
Today I finished my first PCB of the amplifier, I am showing a video of my setup, loading the amplifier 2 Ohms, 1Khz Sine wave, using my UNREGULATED SMPS to drive the amplifier.
I really like the performance of this amplifier.
Soon I will assemble the other PCB with for the +-80V version
Some pictures and the video, ( the music was too loud in the video, sorry for that!)
https://www.youtube.com/watch?v=4T6JIH9cUzo
More Info
diysmps
I really like the performance of this amplifier.
Soon I will assemble the other PCB with for the +-80V version
Some pictures and the video, ( the music was too loud in the video, sorry for that!)
https://www.youtube.com/watch?v=4T6JIH9cUzo
More Info
diysmps
Attachments
Last edited: