Upgrading Marchand XM1 active crossover

[kimbo]

I'm just about to drop in some OPA 2134s into my XM1 to replace the stock LF353 s and will also probably follow Phil Marchands suggestion to "..... replace the 2K trimpot with a fixed metal film resistor.'

I have just made a post to an old thread "OPA2132 bypass" ...

OPA2132 bypass

where Carlosfm and other discuss using bypass caps and there are several other posts that have somewhat heated discussions about this topic, but then realised that it was in the chipamp forum so may not get the attention of other XM1 users.

I'll probably go ahead and add a bypass cap between the +V & -V pins of each chip.

So my question is...."Has anyone else out there upgraded their XM1 and would like to share their experience?"

[Magnetar]

Use 4562 opamp, ceramic bypass caps, teflon filter caps and non metal film matched resistors, it will sing

[kimbo]

Thanks Magnetar for your prompt reply. I hope you don't mind my flurry of questions.....I'm keen to learn.

I just did a quick search and found your previous comment regarding the 4562 in your own crossovers Sensitive Wide Range Open Baffles

Wish I had seen it earlier........however I already have the OPA2134s so I will be running with them for the time being....but you know how these things go.....eventually I'll probably want to try and improve again and will consider your suggestion for the 4562.

I'm thinking that the opa2134 are not quite as fast so the use of bypass caps won't be as critical but I'll give it a go anyway.

I hadn't considered replacing the filter caps......but maybe I will. there is a 10uf electro on each Xover board so I imagine this is what you were refering to? Or did you also mean the main PS electros....2 x 1000uf and 2 x 10 uf? (note the PS is regulated)

And your reference to 'non metal film matched resistors'.... I presume you are suggesting that it is very important to match the 4 resistors on the DIL plug in board that determine the Xfreq.....or are you suggesting that I replace all ten or so resistors on each main Xover board?

By 'non metal film' are you refering to carbon film or ....? Perhaps an example of a brand would make this clearer to me.

Any Aussie members out their know of a good local source for teflon caps ( or a reliable OS one)?

[Magnetar]

Originally posted by kimbo
Thanks Magnetar for your prompt reply. I hope you don't mind my flurry of questions.....I'm keen to learn.

I just did a quick search and found your previous comment regarding the 4562 in your own crossovers Sensitive Wide Range Open Baffles

Wish I had seen it earlier........however I already have the OPA2134s so I will be running with them for the time being....but you know how these things go.....eventually I'll probably want to try and improve again and will consider your suggestion for the 4562.

I'm thinking that the opa2134 are not quite as fast so the use of bypass caps won't be as critical but I'll give it a go anyway.

I hadn't considered replacing the filter caps......but maybe I will. there is a 10uf electro on each Xover board so I imagine this is what you were refering to? Or did you also mean the main PS electros....2 x 1000uf and 2 x 10 uf? (note the PS is regulated)

And your reference to 'non metal film matched resistors'.... I presume you are suggesting that it is very important to match the 4 resistors on the DIL plug in board that determine the Xfreq.....or are you suggesting that I replace all ten or so resistors on each main Xover board?

By 'non metal film' are you refering to carbon film or ....? Perhaps an example of a brand would make this clearer to me.

Any Aussie members out their know of a good local source for teflon caps ( or a reliable OS one)?

The opa2134 is a little better than the lf352. Worth putting them in. Mainly quieter and less fatigue. They also seem sweeter in the top end specially when hopped into class A. The nationals are better in all respects - the bass is much better defined and the overall sound is more more real and transparent.

The filter caps are the little filter caps in the signal path. Probably 3000 nf or 3300 nf - try the russian teflons on ebay - bigger upgrade then opamps ime

Riken and allen bradley matched carbon resistors sound better. Metal film to me is step into fatigue and grain where the carbons are more natural less 'glassy' sounding/

[kimbo]

thanks again magnetar......I'll check out all your suggestions and see what I can source locally or via ebay.

[audiosteve]

Just an FYI,

I had some XM-1's years ago and measured them in our lab at work. I found ots of crosstalk between the HP and LP sections. It was years ago but if I remember correctly the looked much more like a low Q notch filters than a real hi pass or lo pass. I think there was a reason Marchand developed the XM-9. Of course this is all from memory over ten years ago. Your mileage may vary.

[kimbo]

this may be totally unrelated, but I did find when I first built and tested them a year or two ago, that there appeared to be some crosstalk and a distorted signal was appearing on the high pass when compared on the scope. However after chatting to Phil he suggested that the sin wave my signal generator was puting out was not a pure sin wave.....therefore the xover was separating the high and low components of the signal.

As he put it....

"When you monitor the high pass output with the frequency set below the crossover points, the fundamental of the signal gets removed, and you mainly see the distortion products ot the input signal."

I later found that he was correct and my el-cheapo signal generator (kit built) was really just a toy.

However it sounds like you have access to real lab quality test gear? So perhaps this is showing up a real problem with the units.

unfortunately I'm not really in a position at the moment to do more than upgrade a few components to try and maximize the performance of what i have........perhaps in a couple of more years I'll build up credit for chores around the house to use against taking up the kitchen table for a week whilst I work up something better!

[Davey]

The XM-9 is essentially the same thing as the XM-1 except there are dedicated buffered outputs added vice the direct pickoffs from the integrators on the XM-1. The basic state-variable filter circuitry is essentially identical for both units.

I'm not sure why you measured crosstalk between HP and LP sections, but it was probably a XM-1 circuit that was operating incorrectly or something caused by your test configuration.

Obviously if you input a signal within the passband of one filter you'll see a reduced version of it in the other output. That's not crosstalk, that's just the normal filtering action.

Both the XM-1 and XM-9 have limitations. Asymmetric slopes are not possible, and different turnover frequencies for hi and low pass sections are not possible. This is inherent to the state-variable topology.

The XM-44 is a much more versatile unit.

Cheers,

Dave.

[audiosteve]

Look at the board layout of the XM-1 vs. XM-9. The layout and grounding of the XM-9 is FAR superior. I ran these tests years ago on a network analyzer that did a frequency response test from 20Hz to 20 Khz. I don't think I did anything wrong and the test was simple and the results easy to interpret. The same test on the XM-9 showed very clean results. Having said all that, it was ten years ago, I no longer have the XM-1's nor access to the network analyzer. So I can't go back and verify any of this. The only point I was trying to make is that, based on my experience, the XM-1 seems to be an inferior design, so why waste time throwing good money after bad when better options, including the XM-9, are available?

[Davey]

Well, Kimbo already has the XM-1 and he's trying to upgrade it. Maybe it's an inferior design, maybe not, but that's beside the point.

I used/tested both the XM-1 and XM-9 crossover boards. They both work fine if used correctly.

Op-amp rolling in a design like this is hard to gauge without some on-the-bench testing. The op-amps are all configured as integrators vice a "normal" filtering configuration. Any good FET-input op-amp should work okay, but whether they will have the audiophile magic is another subject.

Cheers,

Dave.