Just got an HP 8903E -- mods?

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
HI all -- today UPS delivered an 8903E -- the one without the oscillator built-in. I've cleaned it up a little, checked supplies, cleaned board connections, and cal'd it.

With my Cordell SV oscillator, 1kHz THD (30kHz LP filter in) is 0.0017%, and with a 10kHz single-pole LP on the 600 ohm monitor output, my TRMS HP3458 says it drops to 0.0014%. This is good, especially since it is completely automatic; but have any of you modded or upgraded one of these or an 8903B?

Looks like some signal path distortion could be reduced by replacing the JFETs and 5534 at the input with an OPA134, and then replacing the following 5534 diff amp with the LME-whole-bunch-of-numbers. Then ditto with the LME in the notch filter. Any thoughts?

This one has the CCITT and C-message filters -- I'd be happy to donate either one to anyone who can use one or the other, I'll never use both, but I might convert one to a 400Hz HP filter. Or if you have a 400Hz HP and need one of the others, we can swap....
Joined 2009
Paid Member
I was considering doing something similar to my 8903A. I can't decide if it's worth it. HP did a good job on the design of those analyzers to begin with. I think the analyzer is mostly limited by its noise floor, so substituting some of the op-amps in the front end with LME49xxx's would probably make a difference. I'd do the math first, though, to see if it makes sense.

Please keep us posted on your progress if you decide to mod this.

@skidave -- eBay, $350 + shipping. Took a chance, listing essentially had zero info. Paid off this time. Cosmetics are rough, but operation is solid.

I've always been curious about these, given that they are state-variable filter designs -- in the oscillator of the B version too. But high-resolution is not their forte -- ease of use and multiple measurement features make them ideal for production testing of radio receivers and telephone gear, etc. For hi-fi, not so good. But since I've got it, 'm going to see how much can be done to make it better.

I built Bob Cordell's wonderful analyzer, then sold it a while back -- wasn't using it. Now I wish I had it back.
Just a quick update -- this unit had the rear input/output option (001). There was a hole in the front panel for the monitor out, and the input cables easily just loop back to the front panel, so I took the panel apart, marked the existing subpanel BNC connector holes onto the back of the front panel, and drilled them out. Easy conversion to front inputs. Oh, and I only had to add some coax to the monitor out connector's cable to reach back to the rear of the main board where the solder connection is.

This is a great analyzer for anything where you're making adjustments and need to check performance often -- it's completely automatic and gives results in a couple of seconds at most -- very fast. I'm beginning to love this thing, and I hope some opamp upgrades and a few other changes will give it significantly better performance.

IMHO, an 8903B would have it all over a 339A, and these days, they cost about the same....
I have a 8903A, it is a good tool.
I am designing an audio teser for my self, I can not design a good input amp, so I copied it from 8903's schematic. I try to use OPA2604 and OP37 in it. But I have only old HP3456 voltmeter, I'm not sure I could make it better than 8903.
The 3456 was one of the all-time great DMMs and still is. Insanely stable, and just very, very accurate. With the math functions, it's just a wonderful piece of gear. Trust it, and get it cal'd if you can.

As to the 8903 input, yes doing it with 2604 or two OPA134s and followed by an OPA134 or LME49710 should make a great balanced input stage. I'm going to work on the 8903E after I get the HP 334 greening done.
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.