Perhaps it's my speakers then that produce that lively sound - but I don't necessarily think so. IMO, it's anything but tube-esk like sound. Having said that though, the only class D audio amp that I hold in esteem is the class D ICE amp modules that Pioneer uses in their SC-xx HT receiver line. I've said it before and I'll state it again; they are the smoothest amps I've heard this side of the high-end audio - which I've heard very little of. And they continue to best any class D amp I've built. For sure, Pioneer and AirStudios have optimized the sound for HT use, but even in Pioneer's Pure Direct mode (no DSP), I could listen to them for hours without any fatigue whatsoever. It's like smooth butter to your ears.
In fairness, my IRS2092 amp was connected to bookshelf speakers, so I wasn't ever listening at reference levels or trying to fill a large room with big sound. But those same, simple speakers sounded marvelous with the ICE amps. I say was, only because I don't currently have it hooked up to any system, since I just moved and in the process of building out my listening/mancave room.
Some people scoff at the little LJM 2092 modules, but for a budget amp I'm quite happy with the way they sound. I've never had one hooked up to a sub-woofer either, so that would make a big difference in the amount of bass I hear from my garden variety bookshelf speakers. To each his own I guess.
Could be an impedance mismatch I mean not your speaker's fault per se. Over at Audiocircle the Classdaudio.com 2092 amps compete with 4 and 5 figure amps including tubes and that's a picky audiophile bunch - but like I said there were definitely a few who found the amps bright and tipped towards the treble and impedance matching was thought to be a possible reason. Also I do take Cristi's criticisms to heart and the quality of the ebay boards could play a role. A proper 2092 board shouldn't sound just good, it should sound amazing.
I certainly don't dislike the sound of the LJM ISR2092 based-amps. Goodness knows. I have 5 of his design so I'm happy. I'm sure the sound can be improved even more with better speakers and listening arrangement. None of which I have at the moment.
I recently replaced a Sure module with one of Cristi's design and it does sound very, very good. Better than the LJM? I'm not so sure. That's a close call. Only because I've never done any real serious A/B listening/testing of the two.
I use 12V trigger from my Emotiva UMC-1 to power up my A1000SMPS with 5 L15D's . The 12 V from the trigger activates a simple 12V 5A dual pole relay in the Amp's mains supply . Works very well for more then a year now .
Don't forget a diode over the relay coil .
Cheers ,
Rens
Rens,
I'm a bit of a newbie when it comes to electronics, any chance you have a picture of the layout? I've lurked around here for a while and I'm sure once I get the actual electronics I'll understand it better, but for now I'm a bit confused (which is why my first AMP is coming as assembled components). I'm a software engineer by trade.
From redjr:
I certainly don't dislike the sound of the LJM ISR2092 based-amps. Goodness knows. I have 5 of his design so I'm happy. I'm sure the sound can be improved even more with better speakers and listening arrangement. None of which I have at the moment.I recently replaced a Sure module with one of Cristi's design and it does sound very, very good. Better than the LJM? I'm not so sure. That's a close call. Only because I've never done any real serious A/B listening/testing of the two.
I completly agree:
I am not an ingeener, but I think the most important is the result, so maybe we need some KPI or indicators to compare user satisfaction for class D chips and architectures.
Saying that, I don't really understand why there are so many discussions around post filters, etc... For me, this is just different architectures to compare, with a user approach.
For L15D, there are measurements done in another thread that prove the quality of the sound.
Maybe it is not as good as a $1000 amplifier, but the result is more than decent compare to some class B or other class D.
So, what else to discuss?
My system:
I have a LJM L15D with a Connexelectronic SMPS.
a chinese Matisse PREAMP,
a HIFIMEDIY SABRE DAC with a RASPBERRY as source.
I have to say it is not perfect with high frequencies (it is a little bit crispy), and low frequencies are a little bit flat, but the averall sound is really better than my expectations, and I have to say the sound quality is far better than my Denon Home cinema amplifier.
The greater concern is whether the LJM boards on ebay now are performing to spec. From what I have read through this thread it sounds like initially they were fine but now at the very least some are not. When you buy these from ebay it's not like you are buying them from LJM, there is no seal of approval just because the board says 'LJM' so his measurements are irrelevant at this point. What you're hearing could be a poor implementation. If you like what you hear great, but I wouldn't judge the LJM IRS2092 vs other amps unless I knew for sure it was performing to spec.
Last edited:
I'm planning on getting a couple of these modules to put togther an amplifier. I have a 300VA 115v to 38-0-38 toroid transformer to make a power supply with. This will give me an output voltage of around +/-55VDC. Which version is the proper one to go with this? The L15D or the L25D?
Also, this will be my first power supply. I see I need a bridge rectifier and output filter caps. Are any other components required or is it really that simple?
Also, this will be my first power supply. I see I need a bridge rectifier and output filter caps. Are any other components required or is it really that simple?
After rereading some of this thread with my question in mind I see that the L15D can handle this voltage (a few posts said it was for a lower supply voltage and through me off.).
As for the power supply I'm looking at picking up a PCB, capacitors, and rectifier kit to use with my transformer. The two I'm looking at are:
68 220uF board: 68pcs 75V 220uf Original RUBYCON YXF Capacitor Power supply board kit /PSU kit #0506 13-in Other Electronic Components from Electronic Components & Supplies on Aliexpress.com
or this 8x 2200uF board: LJM Audio Power supply kit ( 8 X 63V2200UF capacitors) suitable for L20,L12 2,L6,MX50,Quad 405 amp board-in Other Electronic Components from Electronic Components & Supplies on Aliexpress.com
Any thoughts on which would be best? Or other suggestions?
Thanks.
As for the power supply I'm looking at picking up a PCB, capacitors, and rectifier kit to use with my transformer. The two I'm looking at are:
68 220uF board: 68pcs 75V 220uf Original RUBYCON YXF Capacitor Power supply board kit /PSU kit #0506 13-in Other Electronic Components from Electronic Components & Supplies on Aliexpress.com
or this 8x 2200uF board: LJM Audio Power supply kit ( 8 X 63V2200UF capacitors) suitable for L20,L12 2,L6,MX50,Quad 405 amp board-in Other Electronic Components from Electronic Components & Supplies on Aliexpress.com
Any thoughts on which would be best? Or other suggestions?
Thanks.
Transformer output of 38 VAC RMS will give a peak of 53.7 Volts but then you have to subtract two diode drops.
Look at the rectifier datasheet and find the voltage drop (Vf, "Forward Voltage") for when the diodes are conducting about 5X your maximum RMS speaker current. (i.e. Assume the speaker current is continuously going. Then assume that the rectifier diodes only conduct for a short time during each cycle, maybe 1/5th of the time. The diodes have to send all of the current, but in 1/5th [or whatever it is] of the time, since they only conduct during the charging pulses, which might be 1/3rd to 1/5th of the time.)
If the diodes could drop 1 Volt each (2 Volts total), for example, then your DC could be as low as 51.7 Volts.
Maybe some bridges use Schottky diodes, which would cut their drop by about half, possibly giving you another Volt or so.
Look at the rectifier datasheet and find the voltage drop (Vf, "Forward Voltage") for when the diodes are conducting about 5X your maximum RMS speaker current. (i.e. Assume the speaker current is continuously going. Then assume that the rectifier diodes only conduct for a short time during each cycle, maybe 1/5th of the time. The diodes have to send all of the current, but in 1/5th [or whatever it is] of the time, since they only conduct during the charging pulses, which might be 1/3rd to 1/5th of the time.)
If the diodes could drop 1 Volt each (2 Volts total), for example, then your DC could be as low as 51.7 Volts.
Maybe some bridges use Schottky diodes, which would cut their drop by about half, possibly giving you another Volt or so.
I really like the idea of lots of smaller paralleled caps.
But you could build your own cap array, very easily, instead, with no etching required:
If you acquire a bare two-sided PCB, you could drill one hole for each cap, and remove a small ring of copper from around each hole (Dremel tool?), on the top side of the PCB. Then insert one cap lead through a hole and solder on bottom, with other lead bent parallel to top copper and soldered to it.
One side of the PCB would be a supply rail and the other side would be ground.
You could divide one side of a PCB by removing a thin line of copper all the way across, and make two supply rails on one PCB.
The large un-broken copper planes would give VERY low inductance (and ESR). For best impedance characteristics, use very thin PCB; maybe 0.5 mm thickness.
Feed from rectifiers to one edge of the PCB and supply your circuit from the opposite edge.
Don't make the array too long and thin; nearer to square is probably better. And use very short and fat wires to connect board output, so you don't destroy the beautiful low inductance.
There are posts you could find, on diyaudio, where Terry Given proposed this and then built one that he measured with his network analyzer.