Conversion of 500MHz TDS744A to 1GHz TDS784A

[nctnico]

You can attach some switches and just try every combination. Perhaps you need a different combination. It is also possible that a monochrome display (I assume the TDS540C has a monochrome display) isn't supported so there is no known model number for your configuration.

[Galen]

Thanks nctnico. I will need a few days to try this as time tied up these days.
The TDS580C published with the FW5.2 as stated in the tek document, so maybe the FW5.0 doesn't have the ID code for TDS580C.
But anyway, I will have a try.

[Alfons]

Here is a plot that shows the frequency response of my Tek744A-784A.
The device triggers cleanly up to over 1.5GHz, the signal also looks quite good, but the frequency response is already absolutely in the basement. The -3dBm is pretty close to 1GHz. The device is not calibrated. First Picture shows 1.5GHz in ET-Sampling, last Pikture shows 1.5GHz in Real-Time-Sampling.

[Galen]

Thanks nctnico. I will need a few days to try this as time tied up these days.
The TDS580C published with the FW5.2 as stated in the tek document, so maybe the FW5.0 doesn't have the ID code for TDS580C.
But anyway, I will have a try.

After I upgraded the firmware, the banner shows 'TDS580C'.  so shorting R1064 is right for TDS580C.

[Galen]

Has anybody tried TDS520C or 724C upgrade?

[madao]

2 ch and 1 Ghz exists, but only color display model. It is called as  TDS782.   (Firmware 5.2e support also  TDS782C. )

I have a   TDS724C aquisition board as good spare part, but i wouldn't try, it is not worth.

[Galen]

Thanks Matt.  I may try it someday.  but not now.

[olle]

Since I just noted that one of the (TDS754/784/...) acquisition board resistor configurations sets the scope to be a TDS794 (2 GHz BW on 1 channel), I wonder if anyone has tried that one out, and if so - did it work or brick the unit?

---
Real programmers: cat > a.out

[Jay_Diddy_B]

Olle and the group,

Welcome to the forum!!

I have just been checking the specification of the TDS794x family, The 794 is 50 only, which means it, (almost certainly) has different input section to the other 700 series family.

I have not seen the inside of the 794 so I don't know for certain.

Regards,
Jay_Diddy_B

[olle]

Thanks a lot for that piece of enlightenment. It means that I'll stay with my 754C-to-784C conversion for now.

 

Real programmers: cat > a.out

[Tantratron]

Thanks nctnico. I will need a few days to try this as time tied up these days.
The TDS580C published with the FW5.2 as stated in the tek document, so maybe the FW5.0 doesn't have the ID code for TDS580C.
But anyway, I will have a try.

After I upgraded the firmware, the banner shows 'TDS580C'.  so shorting R1064 is right for TDS580C.

What 5.x firmware version did you flash to have banner showing TDS580C

[Pete2]

I just got a 754A from Ebay. Paid too much for it because of postage and import VAT but that's the way it is. It is faulty, the infamous relays + dim CRT but at least I can retrofit an LCD if needed.

Anyway I found this topic and was considering the upgrade but I have no way to calibrate it so maybe I'll skip it for now. Also I am looking at a 754C and I assume it can have the 2M option. Is  this true? But I have to have a correct Acq board or can all of the C models support the high memory option? Any other reasons to choose 754C over 754A?

[Tantratron]

Hello Pete2,

Transforming the acquisition board of a 754 into 784 requires removing 4 capacitors and lengthy re-calibration. I assume but could be wrong that if you change faulty attenuator relays will introduce some parasitic capacitance so you'll be obliged to re-calibrate the 754.

In a few words, the TDS754A or TDS784A acquisition board is the same HW as the C series, the only difference concerns logic CPU board. The A runs with 68020 whereas the C serie runs with 68040 so it is just hacking process.

What country do you reside, how does it work with intra-european import tax ?

Albert

[44kgk1lkf6u]

I converted a TDS540D into a TDS580D.  It came with the jumpers R1061, R1062, and R1063.  I removed them and added R1064.  I removed the capacitors.  Then I ran signal path compensation and internal diagnostics.  Now the start-up screen shows TDS580D.  My firmware is v6.1e.  Unlike the TDS540C, I suppose any firmware that comes with revision D will be new enough.

Results.  The fastest time base went from 500 ps / div to 200 ps / div.  The fastest sample rate for a single channel went from 2 GS / s real time and 100 GS / s equivalent time to 4 GS / s real time and 250 GS / s equivalent time.  10% to 90% rise and fall times are about 250 ps at 200 mV / div and about 355 ps at 100 mV / div after the hack as measured with the measure function.  Changing the vertical scale only seems to affect the rise and fall times when it comes with a relay click.  Switching between 50 mV / div and 100 mV / div does not cause a relay to click.  The edges are about the same in those vertical scales.  Likewise the edges are similar at 200 mV / div and 500 mV / div.  To measure these I used an Onsemi NB7VQ14M CML gate to generate the square waves.  I added a 6 dB attenuator so that the waveform fits on the screen at 50 mV / div.  For coarser scales the attenuator did not seem to make a difference in the bandwidth.

[David Hess]

Bandwidth of the 580D and 784D decreases from 1 GHz at 10mV/div to 500 MHz at 1mV/div.

[44kgk1lkf6u]

To be clear, the external attenuation setting was 1x in all the measurments.

I just got out a 20 dB attenuator and measured the bandwidth at larger amplification.  From 100 mV / div down to 10 mV / div, the edges are 355 ps.  At 5 mV / div, it is 400 ps.  I don't have an attenuator that makes the curve fit on the screen at 2 mV / div.

Table 1-16 in the service manual lists the rise time for the 580D and 784D as 400 ps at 10 mV / div to 1 V / div and 530 ps at 5 mV / div.  So maybe the bandwidth is even higher from 200 mV / div to 1 V / div, but they did not want to guarantee it.

[Tantratron]

To be clear, the external attenuation setting was 1x in all the measurments.

I just got out a 20 dB attenuator and measured the bandwidth at larger amplification.  From 100 mV / div down to 10 mV / div, the edges are 355 ps.  At 5 mV / div, it is 400 ps.  I don't have an attenuator that makes the curve fit on the screen at 2 mV / div.

Table 1-16 in the service manual lists the rise time for the 580D and 784D as 400 ps at 10 mV / div to 1 V / div and 530 ps at 5 mV / div.  So maybe the bandwidth is even higher from 200 mV / div to 1 V / div, but they did not want to guarantee it.

Sometimes it is strange with these acquisition board.

Two years ago, I've purchased a TDS784C from eBay but when I open to check, to clean the all unit then I saw the 4 capacitors so I was not happy after the Belgium seller (the 4 cap should not be there for 1 GHz). However I connected on each channel my Leo Bodnar pulser and to my surprise, the bandwidth was really as a TDS784C.

Go figure out... tekronix left the 4 capacitors normally found if TDS754C but declared as TDS784C

[Rafiki]

I've made the conversion too. Removing the jumper resistor and the four 3p3 caps and I get the attached slew rate with a real world signal and 1GHz probes. The upper trace CH2 is with TEK P6243. The lower one, CH3, looks very very similar when switched to the same vertical resolution but is taken with an R&S ZD10 via adapter. I know the same signal with same ZD10 on another 4GHz scope shows a slew rate of about 300ps instead the 500ps here on the TDS744/784. So the limit here is the scope. Anyway I think this is quite good for this old beast.

Measuring with an HF generator shows that my TDS744/784 just reaches the 1GHz. CH2 has a bit less than 3dB attenuation at 1GHz while the others do have slightly over 3dB / 1GHz. Didn't note the values before upgrade though.

[dxl]

To be clear, the external attenuation setting was 1x in all the measurments.

I just got out a 20 dB attenuator and measured the bandwidth at larger amplification.  From 100 mV / div down to 10 mV / div, the edges are 355 ps.  At 5 mV / div, it is 400 ps.  I don't have an attenuator that makes the curve fit on the screen at 2 mV / div.

Table 1-16 in the service manual lists the rise time for the 580D and 784D as 400 ps at 10 mV / div to 1 V / div and 530 ps at 5 mV / div.  So maybe the bandwidth is even higher from 200 mV / div to 1 V / div, but they did not want to guarantee it.

Sometimes it is strange with these acquisition board.

Two years ago, I've purchased a TDS784C from eBay but when I open to check, to clean the all unit then I saw the 4 capacitors so I was not happy after the Belgium seller (the 4 cap should not be there for 1 GHz). However I connected on each channel my Leo Bodnar pulser and to my surprise, the bandwidth was really as a TDS784C.

Go figure out... tekronix left the 4 capacitors normally found if TDS754C but declared as TDS784C

I was in the progress of updating my qemu NI-PCIIA support and was running the Field Adjustment Software. It complained about bandwidth on my TDS784D. Ok, i thought "Attenuator relais have to be renewed, not a big deal". When i opened it up, i noticed that 2 out of the 4 channel had the bandwidth limiting caps installed. However, they measured 0.5pf instead of 3.3pf. I thought "let's remove them, and see whether it's better". Unfortunately the scope was now showing an Acquisition test failure "No HF step found...".

No one ever soldered to the ID resistors, so I assume the capacitors where installed at the factory. There are also no other signs like type/serial sticker which indicate that anyone had been "upgrading" this thing to 784D.

I have no idea why Channel 2 is failing now - there was one odd thing: one of the 8 screws which hold the PCB + Attenuator hybrid together had come lose, and one of the gold contacts in the contact thingy that's between the Attenuators and the PCB had somehow also bent to the neighboar contact. So i'm assuming that either the board or the hybrid was killed by that, and my wiggling with the board finally shorted something. The lose screw and the bent gold contact was one the Channel 2 hybrid.