Yokogawa DL1540C Fan Replacement

[TonyBe]

Hi,

I'm currently working on a replacement of the damaged fan in my Yokogawa DL1540C oscilloscope. After having a short look, I've checked the old fan to be Sanyo DC Petit Ace 25 - a 80mm 3 pin fan (no longer available). So I've ordered an ordenary Noctua 80mm fan to swap the old one out.

Today the new fan arrived and after mounting it to the frame I quickly realized, that the 12V and GND cables in the old fan were swapped. So instead of GND-12V-Speed it was 12V-GND-Speed. I've checked the web to see what kind of standard that might be but wasn't able to find anything about that.. Hm. Maybe some japanish magic.. Anyway.
So I've changed the pins in the new connector and put everything together.

Now the weird problem. When turning on the scope, the new fan spins up to full speed while the scope starts to beep and throws an error: "fan is not spinning, turn off! ".

What is the issue here. When using the old fan, everything is okay. I've measured the speed signal to be very small (~50mV) with 50mV DC offset. But it works.
With the new fan connected, the speed signal is a weird noise signal. Only when adding a external pullup, a clean 12V rect signal which corresponds to the speed (as it should) appears.

I cannot find any information that old fans (roughly 20yrs) had a different speed circuit in them. What is also weird: when adding an external pullup of 10k to the speed signal of the old fan, the signal just shifts up by 100mV without changing the amplitude of the pulses. Is there some unusual current based circuitry to observe that speed signal which maybe resulting in that strange behavior? Does somebody knows this topology or even faced a similiar issue?


I'm curious what you guys may suggest.
Anyway, thank you in advance!

Kind regards
Tony

[fzabkar]

Can you reverse engineer the area of the circuit that monitors the RPM signal?

If you disconnect the RPM pin from the original fan, is there still a 50mV offset? Does this offset appear on the PCB or at the fan?

[nctnico]

You need to use a fan with the same RPM. A Noctua fan with a lower RPM won't do (I tried that with a different Yokogawa scope recently). Your best bet is to look for Sanyo 12V fans with the same current rating and size (including thickness); these likely have the same RPM.

[TonyBe]

@ fzabkar: Might be possible,.. but the scope is not that quickly to take apart and the populated side of the PCB is not visible... I might do this anyway.

@ nctnico: Hm... is that really the point? You're right that the Noctua fan has only half the speed,... 1800 vs. 3400. If that is the case, might be using a little f multiplier to fake the higher RPM 

Thank you so far for the suggestions.. I'll have a look at that and keep you informed.

Kind regards
Tony

[TonyBe]

Hi,

i've tried to validate the suggestion with the RPM,... i've connected an BJT to the Tacho Signal and switched it with a 101 Hz TTL Signal,... reason being, that i've measured the frequency of the "working" old fan and recreated that. Still no succes. I've also played around with this frequency,... nothing changes.

I then put my DMM in series with the Tacho Signal measuring 1mA of DC current. When removing the fan, the Tacho Pin is at 5V... so it seems, that the Fan draws about 1mA of constant current. When measuring the voltage at the Tacho pin with the fan connected, i can measure 100mV DC.

So that means that the internal resistance is 100 Ohm? When measuring the Resistance of the disconnected fan it completely open circuit (>10Meg). I cannot see where this is going,... It seems like the tacho circuitry is broken, but why would it work with the scope?... Hm. Getting more and more strange.

Regards
Tony

[fzabkar]

@TonyBe, the readings you describe are consistent with a PWM controlled fan. I don't know how this allows for speed sensing, though. :-?

See page 7 of the following document:

https://www.sanyodenki.com/archive/document/product/cooling/technical_material_en.pdf

VIH=4.75 to 5.25 V
VIL= 0 to 0.4 V
PWM duty cycle (%) = T1/T ×100
PWM frequency 25 (kHz) = 1/T
Current source (Isource) = 1 mA max. (when control voltage is 0 V)
Current sink (Isink) = 1 mA max. (when control voltage is 5.25 V)
Control terminal voltage = 5.25 V max. (when control terminal is open)
When the control terminal is open, fan speed is the same as when PWM duty cycle is 100%.

[nctnico]

I doubt the original fan is PWM controlled. How would the scope detect the fan isn't running?
Another option is that the original fan used a rotor blocked signal. In this case the yellow wire should go high when the fan is blocked.

[fzabkar]

I doubt the original fan is PWM controlled. How would the scope detect the fan isn't running?
Another option is that the original fan used a rotor blocked signal. In this case the yellow wire should go high when the fan is blocked.

Maybe the low speed sensor configuration on page 6 is more appropriate? If so, then the OP would need to fake this output by grounding the tach pin or leaving it open, or perhaps "borrow" the internal electronics from the original fan?

[nctnico]

I doubt the original fan is PWM controlled. How would the scope detect the fan isn't running?
Another option is that the original fan used a rotor blocked signal. In this case the yellow wire should go high when the fan is blocked.

Maybe the low speed sensor configuration on page 6 is more appropriate? If so, then the OP would need to fake this output by grounding the tach pin or leaving it open, or perhaps "borrow" the internal electronics from the original fan?

The nicest solution would be to modify the fan. I have done that before. Some fans can be taken apart and the yellow wire output function can be modified:
https://www.eevblog.com/forum/testgear/upgrading-mainboard-in-lecroy-dda-3000-(aka-wavepro-7300a)/msg2748842/#msg2748842

[TonyBe]

Hi Folks,

thank you again for the suggestions. It turnes out, that the yellow signal needs to carry a 100mV Signal for the Scope to accept the fan. I simply pulled the Signal up with 20k, and placed a 100R resistor (as calculated by using the old fan). This results in the 100mV and the scope no longer throws the error.
I agree that it might be some blocked sensor, since I cannot see how the scope will tell the exakt speed by a dc voltage...

Anyway. For me, this works.

Again, thank you for the discussion. Really helps to get in contact with people like you!

Kind regards,
Tony

[fzabkar]

Are you sure that you really need a 100mV output from the fan? Could 100mV just be the VCE(sat) of the output transistor?

[TonyBe]

Hi, sorry for the late reply.. has been a busy time.

I've actually tried to only put a pull Up resistor in, but that doesn't work.
Since the new fan does give a tacho signal, there will be no constant Ucesat.. therfore it seems to be necessary to bring up that exakt voltage,.. Since the 100mV caused exactly 1mA of current, this 100R resistor seems to be critical. If you setting up a transistor to give that characteristic, it should be fine as well,..

Anyway: the scope is running for several days now.. so the repair is a succes!

Kind regards
Tony