SparkFun Forums

[teoyongkia]

Hi guys, I'm new in this industry and would like to seek advice on the following:

I have got a laboratory power supply (front panel and a motherboard connected by a Flexible Flat Cable, FFC). I was told that the serial communication between this two panels has a baud rate of 115200bps.

All the components powered off, isolators shorted, with continuity test checked on both ends. I have the contacts of the TX/RX for the front panel and motherboard wired out and connected each to db9 rs232 connected to one laptop each. (PC1 - front panel - (FFC connected) motherboard - PC2)

with null modem for full handshaking configuration on PC1-PC2, i'm able to reach a baudrate of 921600 (max for the hyperterminal), but with the implementation of the power supply in between i can only go up to 19200, which is too low a rate.

I've shortened the cables, twisted them to reduce noise, but I'm still unable to raise the BR (unable as in i will not be able to transmit/receive on the other end).
I've tried to use a digital oscilloscope on the TX for the rising pulse to determine the actual BR but i'm always getting the similar time for a pulse despite increasing the BR on the comport.

Anyone is able to advise me on what should i try to do?
Positive advice is always welcomed:)
thanks guys!

[phalanx]

It almost sounds like you are trying to connect the serial ports of 2 computers, a motherboard, and your power supply all to the same lines (like a bus) which isn't something you can readily do with RS232. RS485 was designed to for this but RS232 is intended for point-to-point connections. Can you better clarify what it is you are trying to do?

Do you have more details on your power supply? If it does support rates over 19.2Kbps, you will have to figure out how to switch it to them otherwise you will have to run your serial link to it at 19.2K.

-Bill

[teoyongkia]

Hi Bill

Thanks for the prompt response.
So the concept was to communicate two PCs with the power supply (consist of a mainboard and a front panel board, connected internally via a FFC) in between. The Power Supply is powered off, so as not to have multiple serial bus, or to avoid bus contention.

yupp the power supply support rate up to 115Kbps.

Thanks once again for your help guys!

[phalanx]

So it appears that you have 2 problems here. First is with only a single RS232 port on your power supply, you can only connect it to one other serial port. There are ways around this but it's outside the scope of what we're dealing with here.

The second problem is the baud rate of your power supply. If the supply supports 115kbps but you are only seeing 19.2kbps from it, you need to figure out how to change its settings so that it uses the faster baud rate. There is a very good chance that it won't change speeds automatically.

-Bill

[teoyongkia]

Hi Bill

Sorry i should have posted this diagram earlier to better illustrate my connection.

So everything is in power off mode, I want to execute the highest BR that i can go with it just based on the connection shown below.
Is it correct?

Thanks again!

[teoyongkia]

Sorry that the cut off pic on the right side... its the PC2, with full handshaking with the PC1 for DTR-DSR and RTS-CTS crossed.

[phalanx]

Is the front panel and the mainboard 2 different serial ports or does the front panel just break out the mainboard serial port?

If they are 4 serial ports strung together in series, it will not work since RS232 transceivers are not designed to be connected in this manner.

-Bill

[teoyongkia]

Hi Bill

the psu has no serial port, but connected via a FFC with a few IRQ. I've only pull out the TX/RX and GND connection and connect it via rs232.
I've made a mistake in my post, only when the chassis ground is not shorted with the rs232 floating gnd, can only the connection works with 19.2Kbps, as per the image.

I used an oscilloscope too to view the channel signal it seems fine too with a sawtooth graph. (isn't it weird to work without the same gnd btw?)

When the chassis ground is connected with the rs232 floating gnd, the signal has gone haywire with triangular waveform and no signal being transmitted on the other side.

grateful for your help Bill!

[phalanx]

the psu has no serial port, but connected via a FFC with a few IRQ.

I'm getting confused with your description of the setup so what exactly does this mean?

Are you saying the PSU doesn't have a serial port because it lacks a DB9 connector? If so, I consider a serial port any bit of hardware that can communicate serially regardless of the physical connector involved. Using my definition, how many serial ports are present here? Does the PSU have 2 (mainboard and front panel) and then 2 PCs are connected for a total of 4? Is the PSU front panel normally connected to the mainboard? Why are you trying to connect 2 PCs? What are you ultimately trying to accomplish here?

In any case, with off the shelf components you can only make a single point-to-point connection with RS232 so your bus setup will not work correctly or reliably.

Lastly, no matter how you change your wiring you will still be communicating with the PSU at 19.2kbps unless you figure out how to reconfigure it for higher speeds.

-Bill

[teoyongkia]

Hi Bill

I think i get what you meant now.
My intent was to test if i can increase the baudrate of the serial communication between the front panel and mainboard of PSU from 115Kbps to higher ones.
My concept was monitor each side of the front panel and mainboard using serial monitoring with 2 PC respectively.
The conduct was to connect it as per the diagram.

So based on the definition of serial communication, thanks for your reminder, the FFC is a form of serial communication.

So to summarize, can i say that i can't monitor that via serial communication as its meant to be for a single point-to-point connection?

Thanks alot Bill for the help thusfar! i will try to figure out the reconfiguration for the PSU.

Teo

[Valen]

Connecting only the RX line of a pc serial port (and GND ofcourse) should not be a problem to monitor what goes across the TX line of the PSU. Though you would need to know the actual baudrate to make sense of the data. With the wrong rate set on the pc you only get garbage or it is filtered out by the serialport hardware. Trial and miss is hardly an effective method to figure out the actual rate used. A (digital memory) oscilloscope or logic analyser would be better suited to analyse the baudrate/bit-period.

2 or more devices sending with their TX line on the same wire is going to cause disruption of each other. That is not going to work. And a short circuit could develop when 2 different levels are forced on the line.

Also are you sure that FF-cable is true RS232 signal levels (+ and - 12 volt or something), or only TTL (positive 0-5 volt).

[phalanx]

Like Valen said, you can connect a TX signal from your PSU to an RX of your PC and not have it affect the line so long as the voltage levels are compatible. I was under the impression that you were trying to use the PC serial ports to actively communicate with the PSU. Since you aren't, you can use 1 PC to monitor the TX line from the mainboard and the other PC to monitor the TX line from the front panel once again assuming the voltage levels are compatible. The handshaking lines at the PCs are not necessary.

If you do figure out how to increase the baud rate of the mainboard, you will also have to figure out how to increase the baud rate of the front panel since it's highly unlikely that it has autobaud capabilities.

-Bill

[teoyongkia]

Hi Valen and Bill

Thanks for helping me out! U two made a similar point on the serial communication and the BR of the PSU.

Do you think even when i power off the PSU, use it as a medium to pass through to reach out for the 2nd serial port of PC2 the connection will be established?
I felt it made sense to do that, just the electronic components are not affecting or whatsoever.

What say you?

-teo

[phalanx]

Maybe? We don't have enough information to answer that for you but my gut feeling is it won't work properly. Since this is RS232, you shouldn't have it connected that way in the first place.

-Bill

[Valen]

Same here. Al you need is a multimeter in resistance setting to measure which pins on one computer side connect to which pins on the other computer's side. Or if that doesn't work, apply level changes on a wire on one end and see if it propagates to the other end.

I still don't understand the wires you have drawn in that schematic. The bypassing flowcontrol lines and disconnected/floating GND makes no sense. And only RX and TX have connection sockets on the frontpanel? Highly unlikely.