Yep. I used to do that in MathCAD some 20+ years ago. That works very well for gaining understanding of how the circuit works and how to compensate it. It doesn't work so well in cases where the transfer function changes dynamically. For example, the LM3886 has different gain as its output approaches the negative rail (and the VAS starts getting into saturation). That tends to be the reason reality and simulation don't always match. That's not modelled in the SPICE/TINA model either, by the way.
Nice!! Yeah. That loop gain/phase kink in the OPA1611 is annoying.
If and when it is working as intended, it probably deserves a thread of its own here. Until then, a big thank you for sharing your knowledge both here and on the Neurochrome site. Very inspiring, enough info to start looking in the right direction, but not so much that all the challenge is gone either.
You're quite welcome. I'm glad you found it useful.
Tom
Yes, I made a rev 2 already to work around some issues in rev 1 - in particular the fact that inverting amps have low impedance (didn't know this
) and rev2 works just fine.Still no way to measure THD+N yet as I worked a bit lately on my electronic load project.
I'm preparing a blog post about the amp now where I'll detail the schematic and show it in action.
Sorry not be able to indicate you wave sources on the net . Certainly you can find these titles on payer streaming websites . I have them on CD but I tried them MP3 highest bit rate on 300B SE , they all sound the same as in wave format, may be less airy , strangled voiced but very clear.
Last edited:
Youtube
I'm starting a YouTube channel and thinking to do a series on designing and simulating an amplifier from scratch. Basically go through the same process I went with this one here.
Tina-TI simulation, part selection, schematic and PCB design, testing etc.
Hobby level since I'm not a processional, ofc.
Would there be any interest in this?
YT Link: YouTube
I'm starting a YouTube channel and thinking to do a series on designing and simulating an amplifier from scratch. Basically go through the same process I went with this one here.
Tina-TI simulation, part selection, schematic and PCB design, testing etc.
Hobby level since I'm not a processional, ofc.
Would there be any interest in this?
YT Link: YouTube
Of course...
Did you built the amplifier already?
Yes, here are some photos:
One channel tested and tortured, the other WIP - missing the bridge rectifier and the output coils.
I'm planning out the case now and I think I'll go for this one:
Aluminium Case with Heatsink 430x311x70mm Silver Front Panel - Audiophonics
It's a big one but the amp is pretty big. I'm going for 2 toroidal transformers of 250VA each.
My next target is to build speakers and a preamp, but first I'll enjoy this a bit.
For intermediary speakers I'll just buy these ones: NUMAN Retrospective 1978 MKII | Numan
Not sure about their quality, but they have a big WAF and I plan to build a pair myself later this year so I consider them temporary.
Can anyone recommend a good sound card to do some measurements?
Depends how much you want to spend. Focusrite Scarlett are indeed very nice and reasonably priced. Focusrite Clarett are even better. My Audient iD4 is good up to about -95dB THD. If you have much higher budget gear from RME, Motu or Universal Audio offers exceptional performance, approaching -115dB THD+N.
Good work...
Can I ask why did you decided to make double-graetz by the psu-module?
No particular reason TBH. In fact I'm running the amp with a modified PSU with one rectifier per channel instead of 2 and everything is fine.
I will modify the schematic to reflect this. Also I moved the bridge rectifier out of the PSU board since I prefer to bolt it to the case to help cool it.
The amp is complete now but I don't have the speakers yet.. I did some tests with a 4Ohm resistor load for the left channel and one crappy 20W marantz speaker for the right channel... I had to be very careful not to blow the speaker
It looks great and it took the better part of the weekend to wire and test everything... I hate working with mains voltage
Thank you all for your support and advice - I finally got my teenager dream come true - to build a good amp from scratch!
Just ordered a Focusrite Scarlett One and I got REW all setup.
My current on-board sound card has a THD+N of around 2% as measured by REW in a loopback, so useless for measuring any decent amp.
Today I'm getting the speakers as well to this weekend I will finally test and enjoy this amp fully.
My current on-board sound card has a THD+N of around 2% as measured by REW in a loopback, so useless for measuring any decent amp.
Today I'm getting the speakers as well to this weekend I will finally test and enjoy this amp fully.
What do you mean? The PCB revision? Here's the latest one jlm-composite-amp - EasyEDA
[Edit] Hmm I think you're referring to the Scarlett Gen - it's 2nd gen. Didn't know there was a 3rd.
So I got a Scarlett 2nd gen - amazing sound card even for day to day use. The fact that I have easily accessible volume controls and inputs/outputs and zero noise in the headphones especially compared to the integrated one - is amazing.
Anyway, this weekend I set out to measure the THD and THD+N but got into some trouble.
I used ARTA on my windows PC and went throught the calibration procedure:
1. Generate a 300Hz signal with the sound card and measure its output using the oscilloscope/voltmeter and enter the value in the box
2. Generate a signal using a siggen into the sound card input and enter the value in the box
I get that these steps should tell ARTA the real output and input levels, all fine so far.
But next I tried the Spectrum Analyzer with loopback (1KHz signal), outputting -3dB and I expected to read back a peak of -3dB at 1KHz, however I got smth lilke -20dB?!
The manual says that I should adjust the windows volume control until I get -3dB:
This is very confusing as I assumed that after the calibration procedure, the loopback test should be very close as ARTA knows how to generate -3dB (real level) and also knows how to interpret the actual input signal value to also get -3dB.
Anyway, I did what the manual told me and got 0,001% THD and 0,36% THD+N for the loopback test which was a bit disappointing but that is not the real issue.
The 2 big issues were:
1. Measuring a BP amp like mine with a sound card will consistently release the magic smoke from R11 (the resistor connecting real ground to signal ground) because I'm shorting one output lead to the input ground... I did this twice and had to disassemble the whole amp to change that resistor. Lesson learned and I moved on to measure just one output lead from one channel at a time.
2. I got really bad results. Got a huge 50Hz component - around -15dB (as reported by ARTA) even if there is zero noise with the amp connected normally (for listening purposes) with the volume all the way to the max. Also got all kinds of harmonics for a total of 10% THD+N regardless of the volume level.
It looks to me like I'm measuring it wrong and injecting all this crap somehow in the system. The 50Hz component was present regardless if I injected 1KHz or not - as soon as I connected the sound card to the amp. I'm suspecting a ground loop was formed or smth...
Forgot to mention that I connected the sound card to one of the output leads using a resistor divider of 10K and 200 Ohm - so a 51x divider. Since the amp can get to 60Vpp I figured I want to drop this to around 1Vpp.
Anyway, I'm stuck now with the measurement. I considered shielded cables and such but it seems unnecessary as the impedance are not that high.
Any advice?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.