MKS 901p

[Rich Feldman]

Very late to this party, I just powered up my first MKS 901P. LED comes on red, then changes to green.

Analog output is +9.45 V whether at atmosphere, or me sucking on the hose, or connected to a vacuum spigot on the wall here.
My guess is that the analog output mode is not at factory default state, it's set to emulate some other transducer model that's pegged at full scale for my three test points.

Sounds like RS-232 communication is needed before more progress.

Did confirm that 901P connector pin 1, which they call Transmit, is putting out -4.5 V. That's a normal RS-232 level, as the device Marks time.
I think it corresponds to RxD in a traditional modem connection. Signal direction is from the DCE device (modem) to the DTE device (terminal, computer, or USB converter). It's called RD at both ends. Would be on pin 2 of a DE9 connector at both ends, right? Thanks!

[Rex Allers]

The LED action you describe is correct.

You didn't actually tell us what pins on the 901p connector you are measuring for the analog.

I have gotten three of these. They all have the 3-row 15-pin D connector (like VGA monitors use). I assume yours does too. I haven't seen one yet that wasn't set up to all defaults when I got it. If yours has been set to some other analog output curve, you'll need to get the serial connection working to issue status read commands or to change a setting. (I assume you knew that.)

Here's a pic of the circuit for an adapter that I wired up to connect to the 901p.

The top two parts are straight out of the manual. Just to be clear, the left pic of the 15-pin connector is looking into the connector on the sensor. I wired it up just like their picture on the right. For power, I connect to a small 12 V SLA battery (12V 7Ah) and I have a switch in the (+) line so I can turn it on and off.

The list at the bottom probably only applies to what I did. I found a VGA cable and cut off one connector so I had a cable with wires on one end to a female 15-pin that plug into the 901p. The colors are what I learned about the cable after ohming all the pins. On a VGA, pin 9 isn't needed so there's no wire for that. I haven't used the sensor relays yet but I figure the Normally Closed of relay 1 I can probably live without.

Here's a pic of the little board I made for this.

For the two pins I provided to measure analog out, they are just like in the circuit diagram, the (+) goes to the 901p pin 5, and the minus to pin 6.

I don't usually use the analog out, but I tried it tonight. I connected a DVM. At atmospheric I saw 8.88 V. On a pump, the serial measurement output said 2 E-2 (20 mTorr) and the voltage was 4.3 V, which matches the default chart in the manual.

To see how your 901p is configured or to change it you'll need to get the serial interface talking. I checked the connections on my adapter board. On the board I've got a 9-pin female connector that plugs into the 9-pin male on the end of a USB-to-Serial adapter cable I am using. So the pins on the female 9-pin of my adapter are wired just as shown. (Pin 1 of the 901p 15-pin is wired to pin 2 of the female 9-pin on my adapter.)

The command to check the analog output emulation is:
@254AO1?;FF
(254 is a broadcast address so should always work.)

Mine returns:
@253ACK30;FF

So is set to '30' which is the factory default.

[Rich Feldman]

Yay! One MKS901P is all better now, after some hand-typed serial communication using Tera Term. That users-manual figure copied by Rex resolves, on close inspection, the proper TD and RD connection to RS-232 port with a DE-9 connector. A voltmeter gave the same answer: when unconnected, each device's serial output line presents a negative voltage.

The magic kiss was to send the command @253FD!ALL;FF (restore factory defaults) at 9600 baud. Then the device became communicative. The non-default setting with which it was received is gone for good.

A trial command, @253TST!ON;FF, made the green LED start blinking as expected.

Some ID and status queries were unsurprising, with one exception.
DT? "LOADLOCK"
MD? "901P"
TEM? "2.93E+1" (degrees C)
T? "O" (both sensors are OK)
BD? "NAK160". (What's up with that? Expected to see the baud setting.)

OK, time to read atmospheric pressure.
PR3? "7.61E+2"
PR4? "7.611E+2" (new challenge: how repeatable and continuous is the significand's fourth digit, in a device with FSR of eight decades? )

Analog voltage output was now 8.88 V, right where it should be. Got it down to 8.62 V (around 400 torr) by mouth sucking. (Probably not the wisest experiment, but the KF-16 flange looked pretty clean, and didn't taste so bad.) Missing is my other MKS901P, already plumbed up to a hose that will fit mechanical vacuum spigots at work.

Thanks for the original tip, and development help!

[Pablo Llaguno]

You are not the first to try to see if the transducer works by mouth sucking, a very sophisticated method indeed

[Bruce Meagher]

Rich, try BR? not BD? to get the baud rate.

[Ameen Aydan]

Hello,

I am very bad at anything that comes to computer programing so I apologize if I sound very inexperienced. I remember Andrew Seltzman posting code on this thread. He said that it is possible to use Matlab as a program to properly interpret the signals. I'm imagining that you use the analog output then hook it up with an Arduino that changes it to a language that the computer can understand. Then, you use the Matlab software that has Andrews code written into it and allow it to display the pressure reading. Is this how you would go about using Matlab as a programming platform? Or more importantly, is this what Andrew did when he used Matlab?

I hope this is a good approach to the general direction of being able to read the pressure off an MKS 901p. Is there anything I have gotten wrong? I would really love to know how to program this to use it but I really don't have that much time in my hands right now.

thank you

[Rich Feldman]

New chapter & question (near bottom) in my tale of two MKS 901P's. Both procured via ebay some time last summer, and talked about here.

Unit 1 (s/n ending with 2835): Bad start. Plumbed it up and saw analog V_out = +9.45 V with no vacuum. Had previous user programmed it to emulate a different device?

Unit 2 (s/n ending with 3803) was first to meet a serial port & terminal program here, because Unit 1 and the KF adapters had been misplaced. This beginner needed a couple of tries to establish interactive communication, including a "restore factory defaults" command which might not have been necessary. Then Unit 2 gave sensible pressure readings and analog output levels for two pressure cases: atmosphere and mouth sucking.

Soon Unit 1 and the plumbing turned up, but six months passed before another serial port session (today).
Now I think Unit 1 might be really broken.
Serial communication worked from the get-go. Baud rate query "BR?" returned 9600. Hourmeter "TIM?" returned 6360.
Transducer status "T?" returned O (both sections OK)
Analog output "AO1?" returned 30 (factory default log function)
but
Pressure reading "PR3?" returned 2.86E+3 torr
Measured voltage was still +9.4 volts.
Those didn't change after factory defaults command "FD!ALL".
Anybody out there familiar with this kind of misbehavior?

Back to Unit 2. "TIM?" 5010 hours. "PR3?" 7.61E+2 torr. Analog V_out = 8.9 volts.
Now it's connected to a facility vacuum port at work, and V_out has been creeping. I observed 7.4 to 7.6 volts (25 to 39 torr), corresponding to sucking power of about 28.9 to 28.3 inches of Hg. A voltage logger is keeping it company for the night.

[Bruce Meagher]

Rich,

I don’t have any insight into your broken 901p, but can you dump of all the parameters? Cut and paste the following commands into your favorite terminal program when you’re connected to the device.

@254PR1?;FF
@254PR2?;FF
@254PR3?;FF
@254PR4?;FF
@254BR?;FF
@254AD?;FF
@254RSD?;FF
@254MD?;FF
@254PN?;FF
@254DT?;FF
@254MF?;FF
@254HV?;FF
@254FV?;FF
@254SN?;FF
@254SW?;FF
@254TIM?;FF
@254TEM?;FF
@254UT?;FF
@254T?;FF
@254U?;FF
@254GT?;FF
@254VAC?;FF
@254ATM?;FF
@254AO1?;FF
@254AO2?;FF

These devices are easy to disassemble, and o-rings are used for sealing. Maybe take yours apart and see if there is any visible corrosion. Below is a pic of one dissembled. Remove the two bolts next to the rs-232 connector then press the two retaining pins on the side to remove the shell. The top board is just held on by the connectors. The bottom board with the two sensors is connected by 4 screws. The piezo sensor is visible on the bottom but the pirani is feed by the little tube and the guts are hidden from inspection. I believe this particular pirani sensor is described in patent US5557972A https://patents.google.com/patent/US5557972A/en

Bruce

[Chris Giles]

BD? "NAK160". (What's up with that? Expected to see the baud setting.)

In case anyone else sees similar, the manual says:

The following list provides descriptions of the NAK codes that may be returned.
NAK Code Error description Example
8 Zero adjustment at too high pressure @253VAC!;FF
9 Atmospheric adjustment at too low pressure @253ATM!7.60;FF
160 Unrecognized message @253S%;FF
169 Invalid argument @253EN1!of;FF
172 Value out of range @253SP1!5.00E+9;FF
175 Command/query character invalid @253FV!;FF
180 Not in setup mode (locked)
...
If the transducer is supplied as a special version (P/N: 901P-xxxxx-xxxx) with pre-configured parameters such as setpoint settings, the setup is per default locked. The transducer will reply with error code “NAK180” if the user tries to change parameters. To change pre-configured parameters refer to unlock procedure page 21.