• Disclaimer: This Vendor's Forum is a paid-for commercial area. Unlike the rest of diyAudio, the Vendor has complete control of what may or may not be posted in this forum. If you wish to discuss technical matters outside the bounds of what is permitted by the Vendor, please use the non-commercial areas of diyAudio to do so.

Which IRS2092 modules?

Status
Not open for further replies.
Noah: It looks like you have spent a good amount of time researching your amp choices before deciding on the Connexelectronic IRS2092 Amp module. Would you mind sharing your thought process?

Below is summary of choices that I have come across and it can be a little overwhelming determining which is best. I guess "best" is subjective, but I would like high quality for a reasonable cost. Below is a summary of what I have come across:

Sure Electronic's "Tripath" Amp- These are low cost, but I see numerous mods to upgrade these amp parts. This does not appeal to me.

HIFIMEDIY- These look like upgraded versions of what Sure Electronic offers and people seem pretty happy with the cost to value relationship.

Hypex UCD 180/400/NCore- This brand look like Class D leader, but carries the highest cost, especially the latest NCore technology. It looks like you can get the "standard" 180/400 amp modules for a reasonable cost. How would these compare to IRS2092 modules you are interested in?

ClassDAudio- There is not a whole lot of information of these, but they appear to be well regarded for less than similar offering from Hypex. I would not expect these to be at the NCore level.

Connexelectronic IRS2092- I really have not heard of these, but I know that Cristi has earned a very good reputation for a quality product at a reasonable price point. Are you expecting these to perform at the ClassDAudio or Hypex level?
 
Hi Mike,

I've read pretty much only good things about the IRS2092-based amps; search for threads with L15D and L20D in the title.

I was looking at the ljm-designed modules on ebay, but can only think parts quality suffers at those prices.

I've read nothing but good things about the build and sound quality of Cristi's amps and PS, and have a much better feeling about spending my money on those.

I thought hard about the ncores.

People are raving about them, but they also go on about cables and such which I certainly can't hear a difference in.

When I finish my amp, I'll see if I can find someone in the area with ncores to compare, and if there's enough difference I can always build an amp with those, likely just for the L/C/R's.
 
IRS2092 alternatives

Noah, You can either wait till the dedicated power supply will be available (hopefully in few weeks) or chose any of the existing suitable sollutions, as explained before, two SMPS800RE would be the most suitable for the moment, except the lack of power cut-off in case of DC component.

If you gonna chose any other vendor products, I can only suggest to chose Hypex nCore. I do not have them or listened any of them, but I trust Bruno's skills and I believe he's the one who can do something really impressive. The next would be only HIFiMeDiy, but I guess IRS2092 still sound better.
Other vendors such as Sure or L15, L20 I wouldn't even bother to consider, even if would cost 1$. They are made of either fake, either dismatled components, see the attached close-up view (and you can search more on the forum from user's posts) where you can clearly see the IRFI transistor is grinded and laserprinted, then the output inductor have no marking, is a cloned version of the original sagami which is a very good inductor. the fake one is geeting hotter because of cheaper core material which exhibit higher losses and poorer linearity, and also much lower saturation current. If the fake inductor and transistors is used, he THD curves copy-pasted from datasheet of IRaudamp 7 have nothing to do with the reality, as the most important components are out of specs. these boards features no over-voltage or under-voltage protections, and no over-temperature protection, which I consider that even the cheapest amplifier ever must feature them otherwise will become useless after first shortcircuit or overtemperature. the output capacitors used is the wrong type, used for EMI filters, and has larger ESL than the correct ones which of course cost more. the board layout is terrible for a class D amplifier, single layer, no copper pour, with huge GND loops which will generate a lot of EMI, and will suddenly blow up when you less expect this due to overvoltage spikes associated with poor layouts.

Note that most ebay sellers do not offer warrany or offer for an extra ~8 usd, no tracking number for shipping, or add few usd, so the overall price if you chose all these option is close to 25-30 usd. At a similar price I can offer this board Connexelectronic
which is much more decently designed and built, use only genuine components, have all the protections which the cheap boards lack of, continuous copper plane on top layer, large output inductor for higher saturation current, offering better linearity, and the size is just a fraction, 35x95mm and 35mm tall.

I have sent some boards for test to a PA company and they will send me the full test report including THD graphs measured with Audio-Precision system two, and as soon as I will get them, I can complete the manual and include the real measurements too, not copy-pasted from IRaudamp manuals, which are totally different than the real one if different parts and layout is used.
 

Attachments

  • overview.JPG
    overview.JPG
    208.7 KB · Views: 1,234
  • fake transistor.JPG
    fake transistor.JPG
    157.9 KB · Views: 1,216
Cristi, when I try to add the modules to my cart, it says "Your Shopping Cart is empty!"

But if I go back to the product page, at the upper right it says

"4 x CxD250-HP
11 x CxD300-8R
1 x CW2C-10A-T"

That's one CxD250 too many, but I don't know how to change it.

If I go to checkout, it says again "Your Shopping Cart is empty!"

Maybe you could just send me a Paypal invoice for

3 x CxD250-HP
11 x CxD300-8R
1 x CW2C-10A-T

to nlkatz a t gmail.com

Thanks
 
In the mean time I can continue to work on the case layout.

The new SMPS board will be 100x150mm, will fit into a 1RU case, and I will give the exact size next week. The connectors will be places at the extremities of the board, for shortest routing path to amplifier modules.

Are the AC in and DC in roughly the same locations as on the SMPS800R?

Also, if they have IEC connectors for AC input like SMPS, could I get instead screw terminal blocks?

Thanks
 
IRS2092 mono amp module - can it run from 89V rails?

Hi Cristi,

I have a linear unregulated power supply with +/-89V rails and need an amp for it. I'm thinking of one of these (the 600W version):
Connexelectronic

In another thread, you mentioned that IRS2092 modules could go up to 94VDC and be OK.

I am concerned about bus pumping - these will be used for subwoofers. I will be using 8 ohms nominal loads. Let's say I use large caps (100,000 uF or more total) and CRC filtering to reduce rail ripple. Will this reduce concerns about bus pumping?

-Charlie
 
Dear fellows, I couldn't access the forum for more than one week. some major event will take part over here this week and next week and I thing it has to do with connection problems.
The position of mains and output connectors will be approximately the same as the SMPS800R(E) and no IEC connector this time.
Potentiometers for CxD300 must be 5-10K because the low input impedance, and I strongly suggest to drive these modules from a low-impendance source.
By the end of the week I will send you some preview.
 
Charlie, +-89V might be working but if you gonna use it for subwoofer, the OVP might mute the module earlier. of course it can be increased to +-95 to +-98V or so, but you're close to the maximum working voltage of output transistors and is not reliable for long term use.
adding such a large capacitance might affect the power supply, either smps or mains transformer. 100mF at 100V are huge. no need to use such large caps, even to avoid pumping. up to 10mF would be enough. the main issue is not only the bus pumping when supply a class D module with such high voltage, but the fact that is very possible that durring nightime the mains voltage to increase few V so does the DC voltage on the amplifier supply, and will render the amp unusable.
 
Charlie, +-89V might be working but if you gonna use it for subwoofer, the OVP might mute the module earlier. of course it can be increased to +-95 to +-98V or so, but you're close to the maximum working voltage of output transistors and is not reliable for long term use.
adding such a large capacitance might affect the power supply, either smps or mains transformer. 100mF at 100V are huge. no need to use such large caps, even to avoid pumping. up to 10mF would be enough. the main issue is not only the bus pumping when supply a class D module with such high voltage, but the fact that is very possible that durring nightime the mains voltage to increase few V so does the DC voltage on the amplifier supply, and will render the amp unusable.

Cristi, Thanks for your comments and thoughts above. It is good to keep these things in mind. I am following up, below:

I measured my line voltage. It's a bit high for the USA, at 122VAC and doesn't seem to vary much day-to-day, at night, etc. I used the measured AC voltage to calculate the transformer secondary voltage I will get, then accounted for losses and rectification and was conservative in arriving at the 89VAC figure. But I will double check these values.

I disagree on the size of the caps. 10,000uF (per rail, meaning 5000uF in series across the full +/- supply) is IMHO way too small because the caps will have too little ability to deliver power or adsorb power, ripple will be very high and regulation low. I plan to use a linear power supply (not smps) and when the amp is delivering power at frequencies less than twice the line frequency the effective rectification factor in the AC-to-DC conversion goes down by 10-15% and you need to rely more on the caps to supply instantaneous power. Maybe 100,000uF per rail is overkill, however...

How do I request from you modules with the higher OVP limit and rail voltage range when ordering? I assume that this requires changing caps and resistors for higher the voltage level and increased dissipation...

I want to mount the modules on a larger heatsink than what is provided. I do not want to use a fan. Looking at the pair that I already have in hand (the 400W version) it looks like I can remove the heat sink from the assembled modules and re-install it on a new heatsink without desoldering anything. Is that correct?

-Charlie
 
Last edited:
Status
Not open for further replies.