Amazon

Scopes have improved, except for samplers. AFAIK you can't buy anything equ ivalent to the Tek 11800 series new at any price.

My TDS 694C (beautifully refurbished) runs 10 Gs/s simultaneously on all fo ur channels. You can get lots faster ones with more sample memory now, but they're the price of a house on a lake. Mine was $2700. (I paid a bit over the odds but I'm very happy with it.)

Spectrum analyzers, on the other hand, have mostly gone the SDR route and s o deteriorated. They have more features and scan faster, but at a horrendou s price in phase noise.

And the bottom line is that if I'd bought new stuff, I My lab wouldn't have a tenth of the capability that it does.

Cheers

Phil Hobbs

On Sun, 29 Nov 2015 04:58:50 -0800 (PST), Phil Hobbs Gave us:

The Tek MDO family is pretty nice.

For some purposes, I agree, at least for the 4000 series. The cross-couplin g possible between domains is especially powerful--you can trigger off a sp ectral peak, for instance. Sweet. Some of that I can do with the HP 8566B, e.g. I often use it in zero-span mode to watch some peak's height change wh ile I tune stuff. (I was just doing that on Friday for the PMT intermod wor k I've been talking about.)

The 3000 series is much more like a PC with scope, SA, and logic cards--the y're basically separate, so their only real advantages are portability (a b iggie) and bench space.

The main issue with the MDOs is the digitizing front end, which is unavoida bly jittery compared with good YIG-tuned oscillators and Schottky mixers. M y HP8566B has about 30-dB better close-in phase noise than the MDOs, and co st $1500. The tradeoff is slower sweep times at narrow bandwidths, but for me it's better to wait--at least I get there eventually, which I often woul dn't with a SDR-style analyzer.

If I were doing field service, I'd far prefer the Tek, but I'm not.

Cheers

Phil Hobbs

I've been wanting to buy a TDR for some time but can't justify $100k and the PTB don't want to buy from eBay. There isn't much out there.

On Sun, 29 Nov 2015 09:55:19 -0500, krw Gave us:

There are ebay seller that are up in Oregon, right next to you know who. They are selling last years production units still on the shelf at guess you know where. Same models as are still in production.

Got a brand new MDO4104-6 for less than $6k that is now listed for $22k. (that's the dash 6, 6GHz model)

Got a 'used' (must have been a non-powered display) analog CRT scope that was their top model in the day, and it arrived never used, as in brand spanking new at 1/4th the price. The analogs still have a purpose. It is a 2465B 400MHz 4 ch.

I also have in the lab

a Tek DPO 4104 a Tek RSA 280 Real Time spectrum analyzer. a Tek CSA 7404 a Tek AFG 2252 Arbitrary Function Generator a Tek TLA 5202 Logic Analyzer A Tek 2024 scope a Tek 2024B scope a Tek 2024C scope

an HP ESG D3000A Signal Generator

An Agilent N1996A Spectrum Analyzer An Agilent N9310A 3GHz RF signal generator

Seems that almost of that comes from the DDS - whether it's the clock multiplier PLL or the phase accumulator count ratio being a jittery value (skewed propagation delays over 24 bits with no sync latch?).

I've seen a patent (not in my current collection and a quick search hasn't turned it up) where two or more DDSs were mixed, each one dealing with "nice numbers" to give low spurs and jitter.

Sometimes that doesn't work very well - the HP 8510 network analyzer ran Virtex on a 68k. Every little bit of the system had a serial number that had to be registered with the o/s or the whole thing would drop into demo/limp home mode. Once a model dropped off support there was no mechanism to bring it all back together after a module swap repair.

You can get them from dealers e.g. Tucker. They'll probably sell you a serv ice contract if you ask.

Cheers

Phil Hobbs

My list is at

Cheers

Phil Hobbs

Tek 11801/SD24 combos can be had for under $2K. If not ebay, maybe buy from a broker?

Picoscope has a USB TDR for something like $12K. I don't know how good it is.

I did a PCB layout for a super-cheap TDR a few months back, as part of a multi-experiment PC board, but I haven't got around to stuffing the TDR part of the board and seeing how it works. A TDR in the 50-100 ps sort of time range shouldn't be difficult or expensive.

We had Myropcb make the boards, and the pads tend to fall off, another disincentive to play with this.

If you can design this sort of equipment, you can certainly fix it. I don' t get where you're coming from at all.

I basically "bootstrapped" my lab by fixing/flipping broken HP and Tek gear on eBay. The days when you could do that are over, though. When I bought a nonfunctional 8566B for $1500 and sold it in perfect working condition f or the same price, I realized that particular party was over.

-- john, KE5FX

I tried. It turned out to be not worth the trouble.

1. Some equipment required special fixtures to get access to the circuits to troubleshoot. For example, the readout character generator on the 7854 mainframe.
2. Some failures were in sealed modules that were not available. For example, the output amplifier on a 7A29 1GHz plugin.
3. Tucker wanted a fortune to repair each piece of equipment.
4. The cost of packaging the equipment for cross-county shipment was high.
5. Ebay had come online and the value of used equipment plummeted by a factor of 10.
6. At the time, it simply was not worth repairing the equipment. I sold it to a local dealer who had repair facilities.

I agree, eBay devastated the used equipment market. But my main point is there was a very high failure rate on equipment that was bought new. And I'm now wondering if the same failure rate can be expected on old used equipment bought on eBay.

I've found the failure rate of classical (1980-2000ish) HP equipment to be very low indeed. When I left IBM, I still had most of the stuff I'd bought over the previous 20 years, and the rest I gave away. I've also had about

Cheers

Phil Hobbs

It's not clear how much of that you bought new and how much on eBay, so it's difficult to make any concluson.

I only lost a few pieces of HP equipment. One major disappointment was an

My main complaint was with Tektronix scopes. They worked great for a while then developed all kinds of random failures. 2465's, 7104's,

About half and half. I've more or less reproduced my IBM lab (where it was all bought new except for an 11801C). I don't have the the super-expensive stuff like my lamented TDS 7704, microwave synthesizer, and tunable picosecond laser system (which I wouldn't have space for anyway).

At IBM I used to have, among other things

TDS 7704 7 GHz 20 Gs/s scope

TDS 684C 1 GHz / 5 Gs/s 4 channels

HP 71000 series modular spectrum analyzer and tracking generator

HP 3562A two-channel dynamic signal analyzer HP 35660A ditto HP 5145B semiconductor parameter analyzer (floppy drive had to be replaced) Several HP synthesizers (three 3325B, two 1.1 GHz ones I forget, plus a microwave 85660D plus plugins) HP 5372A Frequency and time interval analyzer

Vector voltmeter (ancient) HP power meters and AC voltmeters

A bunch of lock-ins (SRS 830 and older EG&G PARC ones)

Some TEK CAMAC stuff that I never really used much

I build a lot of my own test gear too, but it's mostly specialized stuff like log conformance tests for noise canceller BJTs.

Oh, okay, apart from the 2465s those are all '70s kit, and so a good deal earlier than mine. I haven't had a 7xxx analog scope since I was a grad student. The TDS ones and the 118xx are bulletproof so far.

I got lucky with my 2465B at IBM and (so far) my own 2467. The horizontal deflection hybrid is their weak point IIRC.

Cheers

Phil Hobbs

Is your 2467 the B model with the microchannel plate?

I still have two 2467B's and a plain 2467. The 2467 is running fine but the battery-backed ram on the 2467B's has died and they won't boot. I believe John Miles posted a generic version of the firmware so it might be possible to load it on new ram chips and get them running again. Then sell them - the microchannel timeout is way too quick and you have to keep resetting it which is a pain.

It is also possible to remove the input low pass filter on the Rigol DS

There is a problem with leaking electrolytics but I forget which model. If you let me know which one you have I'll look it up and let you know if the caps need to be replaced.

It's the plain model, but does have the MCP, and the short timeout.

But it's so pretty. And the triggering is beautiful.

Thanks. It's Model 2467, Option 11, S/N B010835.

Cheers

Phil Hobbs

The 246xB units above serial B050000 have a controller board that has four SMT electrolytic caps that leak electrolyte and cause damage to the A5 PC board. It is important to change out these caps and clean up the board to prevent problems you describe. At the same time, you should replace the Dallas NVRAM memory with a socket and a new memory chip. The DS1225Y chip can be copied to the new chip to save the cal constants. If the chip has died, the new chip will need the scope to be recalibrated. That will require some special equipment to go through the full calibration and it take about four hours if you have done it before.

Really bad cases of the leaking caps will require some trouble shooting and a few parts will need to be replaced.

Regards

I bought a tube of those before they went away. What's the intelligent method for cloning it?

Thanks

Phil Hobbs

If you have a good vacuum desoldering system, remove the old Dallas NVRAM and make a copy of its contents. Many eprom programmers will support the DS1225Y chip. Then install a high quality machined pin low profile socket. Burn the contents into the new memory and plug it in.

If you don't have a programmer that supports the memory, email me and I'll help you out.

The correct new part is the DS1225AD and Mouser has them fresh and new. Should work for the rest of your life :).