[ 
Account banned, as it is the return of a previously banned account. ]
Hey guys,
A week ago I found this on eBay - https://www.ebay.com/itm/334842335447
I was wondering if anyone else had already tried that kit, but couldn't find anything.
So I ordered it and it came 3 days later. The seller also offers a "PCB version only" but I skipped it for now.
As other composites that I have this one also needs +/-15.0V for the op amp. The eBay seller offers such a PSU for a reasonable price so I got it as well.
I put up a setup and made some initial measurements. Picture 1 and 2.
Output power around 65W@8 Ohms PSU -> +/-34.0V. Picture 3
Output DC (input shorted) - 1.1mV - not bad at all. Hooked to a speaker it is dead silent.
The composite is stable and does not show any signs of oscillation though it lacks a Zobel network.
I tried to measure the THD using my Scarlett 2i2 (4th gen) but the threshold of -106dB was too high. Sounds very promising 🙂
Picture 4 Scarlett 2i2 THD and picture 5 composite THD @ 1W measured with Scarlett 2i2.
I'm going to dig out my notch filter and low distortion 1kHz oscillator to finish that job.
So more measurements are coming.
Account banned, as it is the return of a previously banned account. ]Hey guys,
A week ago I found this on eBay - https://www.ebay.com/itm/334842335447
I was wondering if anyone else had already tried that kit, but couldn't find anything.
So I ordered it and it came 3 days later. The seller also offers a "PCB version only" but I skipped it for now.
As other composites that I have this one also needs +/-15.0V for the op amp. The eBay seller offers such a PSU for a reasonable price so I got it as well.
I put up a setup and made some initial measurements. Picture 1 and 2.
Output power around 65W@8 Ohms PSU -> +/-34.0V. Picture 3
Output DC (input shorted) - 1.1mV - not bad at all. Hooked to a speaker it is dead silent.
The composite is stable and does not show any signs of oscillation though it lacks a Zobel network.
I tried to measure the THD using my Scarlett 2i2 (4th gen) but the threshold of -106dB was too high. Sounds very promising 🙂
Picture 4 Scarlett 2i2 THD and picture 5 composite THD @ 1W measured with Scarlett 2i2.
I'm going to dig out my notch filter and low distortion 1kHz oscillator to finish that job.
So more measurements are coming.
Attachments
Last edited by a moderator:
It doesn't look like they followed National Semiconductor/TI's recommendations for decoupling. The data sheet says 470 uF || 10 uF || 0.1 uF as a minimum. I've seen the LM3886 oscillate with a 4 Ω load when the output approaches the supply rail if the 10 uF decoupling cap is missing.
I take it a step further than National's recommendations: https://neurochrome.com/pages/supply-decoupling
Tom
I take it a step further than National's recommendations: https://neurochrome.com/pages/supply-decoupling
Tom
When the power chip is included in the composite the whole picture changes a lot as the open loop gain is very, very different and the whole thing behaves very differently.
I'm going to check with 4 ohms load to see what will happen but with 8 Ohms load it is rock stable.
I'm going to check with 4 ohms load to see what will happen but with 8 Ohms load it is rock stable.
Putting the LM3886 in a composite loop does not change the fact that the LM3886 needs good decoupling to remain stable. If the LM3886 goes unstable the composite loop goes unstable too.
The vendor claims the amp runs at ±40 V, so I'd test it at that voltage with a 4 Ω load. Drive it to clipping and see how cleanly it clips and how well it recovers. Composite amps will usually not clip as cleanly as non-composite amps. That's just the nature of the beast. I've seen several composite LM3886 amps get stuck oscillating after clipping, even after the input signal amplitude has been reduced. You definitely don't want that.
Another stress test is to use a square wave on the input. Increase the amplitude until the amp clips and then turn the amplitude back down and see if it recovers. And then, of course, do the same with some capacitance added in parallel with the load resistor. I generally like to see a stable step response even with 8 Ω || 2.2 uF.
Tom
The vendor claims the amp runs at ±40 V, so I'd test it at that voltage with a 4 Ω load. Drive it to clipping and see how cleanly it clips and how well it recovers. Composite amps will usually not clip as cleanly as non-composite amps. That's just the nature of the beast. I've seen several composite LM3886 amps get stuck oscillating after clipping, even after the input signal amplitude has been reduced. You definitely don't want that.
Another stress test is to use a square wave on the input. Increase the amplitude until the amp clips and then turn the amplitude back down and see if it recovers. And then, of course, do the same with some capacitance added in parallel with the load resistor. I generally like to see a stable step response even with 8 Ω || 2.2 uF.
Tom
I measured the output response running 4 Ohms load.
The amp is absolutely stable till 68W (sounds funny, right) supply voltages +/-31.0V.
Above that some minor oscillation occurs, just before clipping. Pictures are coming shortly.
It is definitely not decoupling issue. I tried many different scenarios - 0.1uF/10uF/1000uF, 3.3uF/22uF/1000uF. It didn't change anything.
BUT if the heatsink is cool oscillation does not occur. When the heatsink is hot it is well visible. So it's temperature dependent. I think that maybe using not insulated chip would be better.
I tried square waves as well but nothing extraordinary. Pictures are coming shortly.
The amp is absolutely stable till 68W (sounds funny, right) supply voltages +/-31.0V.
Above that some minor oscillation occurs, just before clipping. Pictures are coming shortly.
It is definitely not decoupling issue. I tried many different scenarios - 0.1uF/10uF/1000uF, 3.3uF/22uF/1000uF. It didn't change anything.
BUT if the heatsink is cool oscillation does not occur. When the heatsink is hot it is well visible. So it's temperature dependent. I think that maybe using not insulated chip would be better.
I tried square waves as well but nothing extraordinary. Pictures are coming shortly.
Hmm... You didn't say whether the picture in Post #1 is of your build and test setup, so I'm going to assume that it is. It's certainly presented that way.
The reason I asked that question is that it looks like your board is the exact board shown in the eBay listing. In addition, it appears to be the exact same setup, i.e., the same wires, same use of heat shrink, etc. That makes me believe that you are in fact the seller of this board and are posing here as a satisfied customer posting about their build.
But I could be wrong, so I ask: Do you have any relation to the eBay seller by any chance? Or any conflicts of interests that we should know about?
If you are indeed the eBay seller or are associated with them then I welcome you to the market. Please keep us posted on how deception works as a business strategy.
Below are the picture of your build from Post #1 and the picture of the board in the eBay listing. All the resistors have the tolerance band facing the same way, except one by the input which is reversed ... on both pictures. The scrape marks on the electrolytic caps are identical and the film caps face the same way. I think it's fair to say that these are of the exact same board.
But you did say that you bought the board fully assembled, so maybe it's just an amazing coincidence that you received the same board as the one shown in the eBay listing. But then I suppose it's also just an amazing coincidence that you just happened to use the same wire colours, wire gauge, and heat shrink colours, placement as in the eBay picture.
This is the picture presented as your setup in Post #1. I've added arrows to show the similarities with the eBay picture.
This is the picture from the eBay listing with arrows added to indicate the similarities.
And this is the eBay listing:
Tom
The reason I asked that question is that it looks like your board is the exact board shown in the eBay listing. In addition, it appears to be the exact same setup, i.e., the same wires, same use of heat shrink, etc. That makes me believe that you are in fact the seller of this board and are posing here as a satisfied customer posting about their build.
But I could be wrong, so I ask: Do you have any relation to the eBay seller by any chance? Or any conflicts of interests that we should know about?
If you are indeed the eBay seller or are associated with them then I welcome you to the market. Please keep us posted on how deception works as a business strategy.
Below are the picture of your build from Post #1 and the picture of the board in the eBay listing. All the resistors have the tolerance band facing the same way, except one by the input which is reversed ... on both pictures. The scrape marks on the electrolytic caps are identical and the film caps face the same way. I think it's fair to say that these are of the exact same board.
But you did say that you bought the board fully assembled, so maybe it's just an amazing coincidence that you received the same board as the one shown in the eBay listing. But then I suppose it's also just an amazing coincidence that you just happened to use the same wire colours, wire gauge, and heat shrink colours, placement as in the eBay picture.
This is the picture presented as your setup in Post #1. I've added arrows to show the similarities with the eBay picture.
This is the picture from the eBay listing with arrows added to indicate the similarities.
And this is the eBay listing:
Tom
Tom, those are no coincidences or similarities. You probably already knew just reading the first post.
Mr. kal.tech (402ebay ) how about an introductory rate for the boards?
And can you sell me a pair of parallel 7293 boards only ? I am 99.99% sure you are same as kaltecs.com
Mr. kal.tech (402ebay ) how about an introductory rate for the boards?
And can you sell me a pair of parallel 7293 boards only ? I am 99.99% sure you are same as kaltecs.com
I figured I would give @sherab the benefit of the doubt. After all, he does write:
If @sherab is indeed the eBay seller then he is engaging in deceptive practices here. I don't know about you, but I don't like being manipulated with, deceived, or lied to.
@sherab: The honest way to start a thread like this is to post about your product in the Vendor's Bazaar and waiting for someone to start a build thread featuring your product. Best of luck with your business.
Tom
