EMU 1820M JRC 2068

[Pentium]

I have a photo of the insides of an EMU1820M that I have here, I like to experiment with an old sound card I have for fun.

http://www.flickr.com/photos/42634417@N00/4717408578/sizes/l/

and want to change the opamps for 2 inputs and 2 outputs to test them out at first and see how they sound before doing the whole I/O. I am wondering which opamps to change. There are 2 inputs on another pcb which relate to mic/line combo's, these inputs suck as I think the line ins go through the preamps which in turn sound as gritty as hell so I am not modding the combo inputs just lines 1L and 1R. Please can anyone assist with identifying which opamps are for inputs 1L / 1R which are the TRS jacks on the right side of the photo.This would help me buy less of the audiophile opamps I plan to use which are about £5 each at the moment. If I was to guess I would replace the 5 JRC2068 which make a T shape in between the two TRS jacks, but this is hardly scientific and someone with more knowledge may understand better.

Thanks

 

[Gus]

I have heard the stock 1820m.  I don't know why you would want to change anything in one.  The preamps surprised me.
Change the parts if you want.  What don't you like about the stock unit?

 

[living sounds]

I had an 1820m once and modded it quite a lot back then. First, you can bridge the output electrolytics, there isn't any DC and the sound will improve right away (better low end especially). The 4562 works in the unit and is an improvement, I've tried it. However, I've found the LME49723 to be the perfect upgrade for the JRC2068 (specs match very well), I've tried it in a variety of units (not the 1820m though), and it's a much cheaper chip than the LM4562, but very neutral sounding. I put a lot of these in my Adam monitors, Lynx Aurora and even guitar pedals, sounds great. Still got a lot brand new left, I could sell them to you for 1,30 EUR/piece + shipping, PM me if you're interested.

Oh, and you need to learn out how to solder/unsolder SMT first, my 1820m was my first try modding SMT, and I destroyed a lot of traces.

 

[Skylar]

I worked for E-MU for the 4 years that the 1820 products were developed, sold, and supported.
The preamp circuit was licensed from Ted Fletcher while the rest of the circuit was designed by the same guy who designed all of the Apogee stuff in the 90s.

I have a stack of 1820ms that sit in my closet as backup in case my primary units ever die (I can't imagine life without them).
My opinion is that you will not be able to improve the circuit by part-swapping.
There may be only a handful of other interfaces ever created that will measure as well as these devices performance-wise.
That being said, performance on paper and performance to your ears are two different things...

I'm with Gus. What about the stock unit do you not like?
They are about as flat / transparent as they come.

 

[living sounds]

They do measure well if you test with sinewaves. But music is something else entirely. And I don't think I've ever seen an Apogee unit use JRC2xxx or JRC4xxx op amps. Another problem is the PSU (which is the computer's) and the clock (generated on the PCI card, transferred via CAT5 cable). For the money it's not a bad converter, but there are other devices measuring worse with sinewaves that sound better. They cost more, of course. But from my experience you really can improve the sound by switching in op amps, I just wouldn't use pricy ones, because of the other limitations. I also found this to be one of the rare cases where a very good external clock (going directly into the half-rack sized unit) improved the sound significantly.

 

[Rossi]

Modern test procedures aren't just isolated sinewaves or sine sweeps but usually a series of multisine signals, which isn't too far from what musical signals are. Also, there's IMD measurements. An in all respects, the Emu 1820m shows excellent performance. My personal impression is that those boxes were very well designed. I just wish the drivers and software were up to the same level.

 

[JoleFIN]

We had an emu 1820m for a year or so and the image I got of it wasn't all flattering. It sounds pretty good for the money, yes, but the low end wasn't as good as on ie. RME adi8ds. So at least skipping the output caps as mentioned above, would prolly enhance this. The reason we only had it for around a year and half was that one day it didn't connect with the computer anymore, and it wasn't about the cable or the connectors. So I'm unsure about the build quality, maybe we had a monday unit.

 

[Pentium]

Thanks for replies chaps, food for thought. Well through listening experience I know the 1820M's DAC section to be a little on the cloudy side (it uses a Cirrus CS4398 chip, fairly decent as I understand). My main DAC/ADC is a much more modern which now has top notch opamps in both L/R outs and headphones and it kills the 1820M dead in terms of sheer clarity, bass, general linearity, top end and most importantly less grain. I only really heard this when I thought the new DAC died on me and I had to listen through the 1820M again, what I heard was not great. The cards spec out excellent and the noise floor literally is non existence to all intents and purposes. I am  up for the fun of tinkering with this. It is very good but I want excellent and hope this will do the trick. You have worried me slightly about the destroyed circuit traces I hope I don't mash it up I might need to practice on something first. MY PC power supply is made by TAGAN which is overspecced in terms of wattage and supposedly very well specced (very stable and rails and has its own chassis earth cable) for a PC powersupply so hopefully it's doing a decent job, the oriuginal CAT5 supplied has ferrite beads at both ends so hopefully this reduces potential RFI from the PC processors (the PSU cost me £110.00 and has worked flawlessly for 6 years).I will proceed with caution, any tips for SMD opamp soldering would be helpful.

 

[living sounds]

Those JRC op amps produce some harshness and grain, but it won't all go away by using LM4562 (which is not the best sounding op amp out there IMO, kind of on the bland/sterile side). For that kind of money I'd suggest getting OPA2211A or rather the lower-cost OPA1612 which has slightly worse DC specs, which doesn't matter here. But again, I'd try the LME49723 first, it's cheap and sounds very good.

The Cirrus chip is great, it's not the problem. A different CAT5 won't change anything, all the signal is transferred via a proprietery clock-free format. A good external clock coming from a converter directly going into the docking stations S/PDIF input will make a difference, I've tried it. But a great converter it will not become. It doesn't change the fact that the switching PC power supply introduces noise, and there's no substitute for a precison clock getting clean power directly clocking the converter (internally).

You can also add 100nf poly bybpass caps to all the other electrolytics in the signal path. The output DC blocking caps are the smaller ones standing close to each other, next to the outputs. I don't remember much more.

But in all honesty, thinking about this some more, I wouldn't advice anyone to mod this card, especially without experience. It's a pain to dissassemble and reassemble, you can easily destroy something and it won't become a high end converter by modding anyway. I learned this the hard way by modding several converters/interfaces and finally gave up and bought an Aurora. Which I then modded later on (changed op amps, made a big difference in there). But the bottom line with converters is that some crucial elements have to be on a high level to start with because you won't be really able to change them on a 4-layer all SMT PCB.

 

[Skylar]

What is this "grain" you guys are talking about?

 

[jackies]

Well, I been replacing opamps in soundcards for years, most recently I replaced 2068 with 4562 in my Echo Audiofire12. Yes, it does sound better. I know there's plenty of people here who dismiss it as scientifically unproved, but sonic difference is there. I suspect the aforementioned "grain" might be intermodulation distortion products. As we all know, DAC produces lots of out of band noise (ultrasonic), and (i imagine) a lesser opamp produces more intermodulation than a better opamp, that's (probably) why some people love video opamps (LT****) even though they are ridiculously overspecced for audio...
That said, I personally did enounter a situation where replacing older opamps (LM833) with 4562 didn't improve the sound drastically - in a RIAA preamp. I wonder, maybe that's because there is no strong ultrasonic content coming from vinyl in the first place?

 

[s2udio]

But in all honesty, thinking about this some more, I wouldn't advice anyone to mod this card,

Good advice...... I use two EMU 1212s.........excellent cards,i really dont think you will improve anything( you may release the magic smoke ?)
And i cant detect this audio " grain " either ??

 

[living sounds]

But in all honesty, thinking about this some more, I wouldn't advice anyone to mod this card,

Good advice...... I use two EMU 1212s.........excellent cards,i really dont think you will improve anything( you may release the magic smoke ?)
And i cant detect this audio " grain " either ??

Well, if you do it right there will be an improvement. As for hearing the "grain" etc. - it becomes obvious when auditioning side-by-side. And there's a big difference recording and mixing through converters, especially with processing involved inbetween. In my own experience, of course.

 

[Kingston]

I know there's plenty of people here who dismiss it as scientifically unproved, but sonic difference is there.

The differences are not actually unproven, and I find it baffling anyone would say otherwise especially when confronted with years of active research. The differences are very clearly presented as several types of distortion and noise (or at the very best the lack of either!). Much of it plotted extensively by Samuel Groner, among others to less extent in the past. The only part left unproven is if these differences matter in studio. And it will always stay that way due to its subjective nature.

I personally have enjoyed much swapping opamps in several "standard duty" positions, especially all kinds of buffers in mixers and AD/DA converters. There are surprising differences to be heard...

 

[Skylar]

I'm just trying to get an understanding of what the "grain" is.
Noise? Distortion?
I definitely don't hear it, but I don't have a setup where I can A/B a modded 1820m versus a stock 1820m.

living sounds, is there a way I could send you one of my 1820ms for modding?
I'm interested in hearing it, but I don't really have SMD experience.

Do the modded units measure differently than stock units?

 

[living sounds]

I'm just trying to get an understanding of what the "grain" is.
Noise? Distortion?
I definitely don't hear it, but I don't have a setup where I can A/B a modded 1820m versus a stock 1820m.

living sounds, is there a way I could send you one of my 1820ms for modding?
I'm interested in hearing it, but I don't really have SMD experience.

Do the modded units measure differently than stock units?

Theoretically, yes, but it doesn't make any economic sense to send this unit half way across the globe and back for modding. There will be a difference, but it's not worth this kind of expense, I think.

Now I can't tell you about the 1820m, as that was too long ago, but with the Aurora the modded channels show better THD specs with sinewave tests. Only a couple db or so, but the sonic difference is MUCH bigger than the one measured by these kind of tests. And the Aurora already has some of the better chips used in the stock version.

About the grain: Besides muddy bass, lack of detail, bad dynamics (compressed sound), impaired transients, the main problem IMO is accumulation of aggressive frequencies in the midrange. Very obvious with vocals that seem piercing and are very had to get to sit in the mix. Sort of like resonances that also stretch a little over time. The opposite is a very relaxed and unobtrusive sound with great detail, clarity, punch, separation. Somewhat like treated vs. untreated room.

 

[Kingston]

send you one of my 1820ms for modding

You're a full on DIY person. De-soldering SMD opamps and soldering a new one in is not difficult or dangerous. Rosin based standard solder-wick and a reasonably low soldering temperature are the needed safety precautions, besides a reasonably steady hand. If you screw up and get a blob covering more than one trace, just solder-wick the thing out. The parts won't break.

Try LM4562 vs. the original first.

 

[Skylar]

True, I just don't have much spare time and am afraid of SMT!
I guess I need to bite the bullet at some point.

Maybe I'll just mod each channel pair of one of my 1820ms differently.
That should provide a good testbed for multiple opamps.

 

[Pentium]

I would probably have been in ignorance unless I heard it in contrast with a better performing DAC, I have always liked the 1820M to me it sounds like a blurryness, lack of definition, loss of clarity and information (less top end detail or simply "duller"). I am listening on $8,000.00 speakers which are fairly high res that may have something to do with it. I don't buy the "there is no difference" between opamps, in many instances I hear it immediately and the differences are not small IMO.To generalize somewhat, OP275's a little bit cloudy compared to the best and a little unexciting but better a top end than 5532 (strong bass)/4558's (harsh upper mids), 4580, harsh, glassy top end, 4562 balanced and accurate all the way through the spectrum thats what my ears tell me and in my line of work thats what I go on. I might be barking up the wrong tree, it could well be clock related. The specs of this card are very commendable but in listening there are better devices at least on the DAC side.Just awaiting the time and courage to proceed !
I might need one of those wicks from somewhere.

 

[Emperor-TK]

I have to disagree with some of the sentiments here about the 1820M.  I've done a direct comparison between the 1820M and a few prosumer and professional converter boxes/interfaces and would put the the 1820M in the the same ballpark as the pro stuff.  To my ears, the difference between the Emu and an Apogee was negligible (but not unnoticeable), while the difference between the Emu and an M-Audio, RME, and my old Ensoniq Paris (highly regarded in it's day) were significant.  In all fairness, I know that Apogee isn't a favorite among some, and I didn't have access to a Lynx, Crane Song, Prism, etc.  I currently have the 1820M, and RME Fireface 800, and a Presonus Central Station and use the 1820M as my main interface/converter.

If I won a $50,000 shopping spree at Mercenary Audio, new converters wouldn't be in my shopping cart.  I have much bigger fish to fry first.

-Chris