Tek 2400 scopes - have the schematics for the modules been released?

[peter-h]

I repaired a 2465B by replacing the trigger module - a common issue - and opened up the faulty module. I posted the internals at the end of this thread

https://www.eevblog.com/forum/testgear/2465bct-error-test-05-fail-44-(24)-trouble/new/#new

There are two chips there. The more curious thing is that the module runs pretty warm (most of the modules in the scope do) so they must be running everything at quite a high current, to achieve the speeds.

They may be Tek custom chips, or maybe not? If they are then obviously the circuits won't come out.

I ask this because I read somewhere of efforts to develop a replacement module. Probably not worth doing now, with so many scopes being parted out and the modules going relatively cheaply.

[David Hess]

Tektronix had their own IC fabrication plant and made their own ICs and hybrids for themselves and others.  Some of their ICs were documented but I have not seen datasheets for any of their hybrids.

[george.b]

I wonder how hard could it be to reverse engineer from die shots. They don't seem terribly densely integrated. It's beyond my current capabilities/knowledge, but wiser people could perhaps figure it out?

[BravoV]

Someone did it, at the popular U800 chip -> Tektronix 2465/2467 series U800 IC replacement design

[David Hess]

For U800 we can make a pretty good inference of its internal design details based on published data for earlier horizontal modules.

[WastelandTek]

Someone did it, at the popular U800 chip -> Tektronix 2465/2467 series U800 IC replacement design

that...is awesome!!

[peter-h]

Impressive! Is he selling the PCB? I see no PCB files in that ZIP.

One could also work out a functional replacement for the modules...

[BravoV]

Impressive! Is he selling the PCB? I see no PCB files in that ZIP.

One could also work out a functional replacement for the modules...

As 2465B owners, few years ago I contacted the author, he mentioned he didn't have any interest to sell them nor providing the board layout.

Also currently I have spare unobtanium U800s at my disposal, hence I lost interest. 

[peter-h]

That is really weird. Go to all that work but not be willing to help anybody afterwards.

It's not hard to design a PCB though - assuming his circuit is working. I see he has no tracks on the top (just a ground plane) and it can't be a single layer board, so 4 layer at least.

U800 gets quite warm. I wonder what in that circuit generates the heat? Nothing specific I can see.

There is a Greek guy, on Ebay as "qservice_rhodes" who has lots of Tek scope parts. He also repairs them for a fixed fee. I bought a trigger module from him.

[David Hess]

That is really weird. Go to all that work but not be willing to help anybody afterwards.

It's not hard to design a PCB though - assuming his circuit is working. I see he has no tracks on the top (just a ground plane) and it can't be a single layer board, so 4 layer at least.

I studied the schematic a couple years ago and the design is very brute force.  He had problems getting sufficient speed out of the current feedback operational amplifiers because the wrong circuit configuration was used and that made stability marginal.  So there is a good chance that a different circuit layout will fail to operate properly.

U800 gets quite warm. I wonder what in that circuit generates the heat? Nothing specific I can see.

The idle current through the high voltage output transistors is pretty high.

[med6753]

Even if you do manage to reverse engineer the chips (unlikely), how are you going to deal with the ceramic substrate that they mount on? Sorry, but I think this is a pipe dream and it won't happen. The newest of these scopes is pushing 30 years old. Who is going to invest time and resources for an obsolete chip and then try to re-create the ceramic substrate and custom heat sink? And an extremely limited market to boot. The ROI would be ridiculous and the part would cost much more than the scope is worth.

Yes, these hybrids run hot. And your best option for keeping them healthy is to adhere to some simple principles. First, the cooling system (fan, etc) must be kept in top notch condition. Second, the ventilation slots...especially on the bottom of the scope....must be kept clear. Third, don't mount the scope on top of another unit that generates a lot of heat. Fourth, it's preferable that the scope be up on it's bail handle to insure adequate airflow underneath.

[peter-h]

The ceramic to PCB socket interface is OK to replicate. You could do a replica of the ceramic with 1.6mm PCB material. The tricky bit would be finding a way to hold the PCB against the spring-loaded pins of the socket.

The latest 2465B was made c. 1996. I have one, with 1995 date code chips, and inside it looks brand new. However, it has a really noisy fan - much noisier than a 2465DVS I also have (just bought it for GBP 250!) and which is perhaps 35 years old.

I agree nobody is going to do this, all the time there are enough donor scopes around to cannibalise and the modules are on Ebay for 50 quid

[med6753]

The latest 2465B was made c. 1996. I have one, with 1995 date code chips, and inside it looks brand new. However, it has a really noisy fan - much noisier than a 2465DVS I also have (just bought it for GBP 250!) and which is perhaps 35 years old.

The 2465DVS uses a Seimens variable speed motor with a squirrel cage fan. It is generally very quiet but has a reputation for dry bearings which results in rattling and squealing. The 2465B Teardown thread in this section has rebuild/repair instructions. The 2465B uses a more conventional, and nosier, computer type fan. 

[peter-h]

Yes I noticed the very nice fan on the DVS. The B has just got a cheap and nasty fan, and I reduced its speed a bit with some series diodes, to make working in the room more comfortable.

These are superb scopes. I wrote this before but at work we have a 3-4k Le Croy which blew up after 1.5 years. Luckily, fixed under warranty, but next time it won't be. And you can pick up a ~20k USD (year 2000) Tek digital scope of similar spec (500MHz, 5gsps) for 1k. Go figure... The Le Croy has a nice big screen and some better datacomms (SPI etc) analysis options.

[WastelandTek]

Yes I noticed the very nice fan on the DVS.

"nice" until you have to unwind it to service the bearings

soooo glad my 2445 seems to have really low hours

[peter-h]

How do you tell the hours?

[WastelandTek]

How do you tell the hours?

Well, honestly, you don't really.

I say "seems" as mine has zero fan bearing noise, was quite clean inside with no evidence of many disassemblies, and was still operating perfectly with its 1983 capacitors.  Also it didn't have any cal stickers or adhesive residue on it that you often see on gear from production labs.

None of this is conclusive, of course, and there is certainly a bit of blind hoping on my part that it was privately owned and rarely used.

[med6753]

The 2445/2465 don't have the "EXER 5" function which records total power on hours. The later A & B versions do.

[peter-h]

How do you get this value to display?

[WastelandTek]

if I remember correctly you hold down delta V and delta T at the same time and then push slope while still holding those down

that gets you into the mode, then you scroll around with the trigger mode toggle, trigger coupling is the enter key

its all in the manual

[peter-h]

I found it. My ~1995 2465B shows 1273 hrs and 329 power cycles. That must be an unusually low time example.

Still, it developed a ribbon cable fault (the cable going to the front panel scanning - I posted it here previously) and the trigger module fault.

[rf-loop]

I found it. My ~1995 2465B shows 1273 hrs and 329 power cycles. That must be an unusually low time example.

Still, it developed a ribbon cable fault (the cable going to the front panel scanning - I posted it here previously) and the trigger module fault.

Some times we can trust these numbers  (power on hours and power cycles) and some times they seem suscipious. If user know his 2465B history then he can estimate if these numbers can trust or not, together with other possible signs about lot of power on hours (looking example trace with some settings it is different if scope (tube) have 1000 or over 60000 hours running behind.)

But it need note that these poweron and powercycle numbers are fully editable. These can write just how want in Calibration process. (Cal 05)  So they are ok only if no one have never changed these.

[peter-h]

FWIW, if somebody did produce a replacement circuit, and if it can fit on a PCB of the same size as the ceramic (about 37mm square), then fitting that PCB into the existing heatsink module case would solve the mechanical issues easily.

The ceramic is about 1mm thick.