Tektronix 2445 teardown pictures

[AlphZeta]

OK. I just got myself a "for parts/not working" Tektronix 2445 cheap. I know a lot of people on this forum had suggested that when sellers state "powers on, no further testing was performed" would most likely mean something was for sure wrong, but according to my own experience, sometimes this simply means seller doesn't have the time or knowledge to test. This is especially truefor those high volume sellers with high feedback scores. That said, don't bet on it either.

Anyway, long story short, the scope I got is actually in perfect working condition and I was able to fix a couple of broken knobs. You can take a look at the teardown pictures on my blog at http://www.kerrywong.com/2013/04/25/tektronix-2445-teardown/.

Like many readers here, I really love these old Tek scopes I think they were so well made and even by today's standard the specs are still pretty amazing.

[w2aew]

Nice write-up and pictures!  Congratulations on the nice scope!

[AlphZeta]

Nice write-up and pictures!  Congratulations on the nice scope!

Thanks @w2aew. These  old scopes were  quite well made, and it's funny how little they go for. The strange thing is that sometimes you can pick up a 2445 on eBay for less than a 2235.

[c4757p]

I approve!  The 2445A is my workhorse scope. I can't imagine parting with it while it is still in working condition.

[nanofrog]

I approve!  The 2445A is my workhorse scope. I can't imagine parting with it while it is still in working condition.

2445B here. 

[casinada]

Nice Pictures
I Have a 2465A DM 1987 Vintage. I love it so much that I just got a 2465B DM to repair . A big order of capacitors is on the way
Nice BLog

[BravoV]

Nice score, congratulation ! 

Quoting you

And in the picture to the right, you can see that a choke with only one pin soldered on the board, the other side is connected to a wire directly. I am not quite sure if these were originally done this way or was modified during servicing afterwards.

Though I'm not sure it was modified or not, but originally they were soldered that way from Tektronix. You can check the similar connection on the A1 main board pic at my 2465B teardown thread -> Here

Now, homework to do ..lol ..  :

• Provide heatsinking on that unobtainium U800 chip.
• Do a major refresh the caps at the switching power supply section and other miscellaneous passive components to prolong its life, a good guide that is applicable for Tek 24XX A/B scope owners since they're all have same power supply boards, download the guide in PDF -> Here
• Replace the old battery backup at the A5 logic board, it holds the cal setting. The price of proper re-calibration might be higher than the price of this scope it self. 

[c4757p]

Do a major refresh the caps at the switching power supply section and other miscellaneous passive components to prolong its life, a good guide that is applicable for Tek 24XX A/B scope owners since they're all have same power supply boards, download the guide in PDF -> Here

Wow, nice! Bookmarked for when I have the money to invest.  Definitely sounds like a good idea. I haven't really done anything with mine other than replace a bad part.

[AlphZeta]

Hey @BravoV,

I am pretty sure 2445 doesn't use a battery to hold the cal info on A5? If I remember the discussions correctly, only 2445A uses a battery?

[BravoV]

Hey @BravoV,

I am pretty sure 2445 doesn't use a battery to hold the cal info on A5? If I remember the discussions correctly, only 2445A uses a battery?

Hmm.. I didn't know that, so how it stores the cal data ? Its not like they had a non volatile memory like eeprom at that era when it was made.

[w2aew]

Looks like there are manual trimpots in the photos.

[c4757p]

2445 has one of those idiotic memory+battery combo packages.

Looks like there are manual trimpots in the photos.

And it's littered with trimpots, even on boards that ought to be all digital, just like the more serviceable 2445A. Yes, it's a doozy of a calibration routine.

[marmad]

Hmm.. I didn't know that, so how it stores the cal data ? Its not like they had a non volatile memory like eeprom at that era when it was made.

Of course they did; I was using the 2816 EEPROM way before your scope was built.

[c4757p]

Of course they did; I was using the 2816 EEPROM way before your scope was built.

I wonder what is more reliable - low power RAM with a long-life backup battery, or an early 80s EEPROM. Seems like they didn't start to trust EEPROMs for calibration data until the 90s.

[marmad]

I wonder what is more reliable - low power RAM with a long-life backup battery, or an early 80s EEPROM. Seems like they didn't start to trust EEPROMs for calibration data until the 90s.

Well, the early EEPROMSs were slow, limited to just thousands of single-byte write-cycles, and quite expensive. They were handy if you couldn't use batteries for design reasons - but otherwise  not worth the cost/effort.

[AlphZeta]

Hey @BravoV,

I am pretty sure 2445 doesn't use a battery to hold the cal info on A5? If I remember the discussions correctly, only 2445A uses a battery?

Hmm.. I didn't know that, so how it stores the cal data ? Its not like they had a non volatile memory like eeprom at that era when it was made.

According to 2445 service manual (section Theory of Operation), unlike 2445A, this scope uses EAROM (electrically-alterable read-only memory), which is a kind of EEPROM to store the calibration constants.

EAROM seemed to be a bit more reliable than the early EPROMs IMO.

[BravoV]

Just curious why Tek used those pesky battery backed up sram at higher and later models like 2465x series instead of this earom ? 

[marmad]

According to 2445 service manual (section Theory of Operation), unlike 2445A, this scope uses EAROM (electrically-alterable read-only memory), which is a kind of EEPROM to store the calibration constants.

EAROM seemed to be a bit more reliable than the early EPROMs IMO.

Yes, EAROM is just a type of EEPROM that is bit-erasable/writable (instead of per byte). Does it list the part number?

[c4757p]

I stand corrected! It does indeed use an ER1400 EAROM! Apparently it's the 2400B series that used a potted battery/memory module (DS1225).

[AlphZeta]

According to 2445 service manual (section Theory of Operation), unlike 2445A, this scope uses EAROM (electrically-alterable read-only memory), which is a kind of EEPROM to store the calibration constants.

EAROM seemed to be a bit more reliable than the early EPROMs IMO.

Yes, EAROM is just a type of EEPROM that is bit-erasable/writable (instead of per byte). Does it list the part number?

The service manual listed it as follows:

Component No: A5U2008
Mfr Part No: 156-1566-00
Descr: MICROCIRCUIT, DI:EPROM, 100x14

A compatible part would be M5G1400P (http://www.datasheetcatalog.org/datasheets/120/398793_DS.pdf). M5G1400P's retention time is only rated for 10 years though.

[marmad]

I stand corrected! It does indeed use an ER1400 EAROM! Apparently it's the 2400B series that used a potted battery/memory module (DS1225).

Hmm... I've had devices fail from lost (early) EEPROM data over time. Are there any documented instances of these devices losing their data? For a long service-life into the future, it might be worth making a little daughterboard with uC that replaces the ER1400.

Edit: I was assuming the part is no longer made - but if it is, you could just read out the stored data and replace it.

[malch]

Just bought a 2445B from Ebay.It should be here after Christmas. The serial number was fuzzy but I think it was xx34xxx.
I will be checking out all the caps and  I dont know if I should be happy if I have a Dallas chip or a ER1400 EAROM chip.
Going to the 2465B tear-down thread.

[sparkybg]

Attached is the ER1400 original datasheet.

Why they started using battery backed SRAM - the question is simple - SRAM is much faster (200ns (now 70ns) vs 10ms write time) 64 kilobits instead of just 100words x 14bits, and comparable guaranteed data retention of both - 10 years.

Today there is a good replacement ferroelectric RAM chips (FM16W08) for the "B" units with 150 years data retention and 10 trillion write/erase cycles. When this chip is used, the problems with NVRAM are over. So IMHO even today the second solution is a better one. Moreover - ths scope saves not just the calibration constants in it (512 bytes out of 8192), but also the last front panel settings and several front panel setups you can recall at any time. It uses it just like ordinary SRAM - it is on the same address/data bus with other RAM chips on the board.

But anyway - there is a separate thread with much more info for the "B" units.