Tektronix Technical Forums are maintained by community involvement. Feel free to post questions or respond to questions by other members. Should you require a time-sensitive answer, please contact your local Tektronix support center here.

2304A Power Consumption Measurement

2302, 2303, 2303-PJ, 2304, 2306, 2306-PJ, 2306-VS, 2308
Post Reply
snoof378
Posts: 1
Joined: June 27th, 2013, 12:36 pm
Country: United States

2304A Power Consumption Measurement

Post by snoof378 » June 27th, 2013, 1:30 pm

I am interfacing with the Keithley 2304A high-speed power supply under linux via GPIB and have successfully communicated with the device and can set the supply voltage, take voltage/current readings, etc. However, I am interested in measuring the average, minimum, and maximum power consumption by a load device over time. With the serial timeout settings I am using (with a serial to GPIB adapter), I can get about 4 current/voltage measurement pairs a second, which is much too slow to characterize the power consumption accurately. Just measuring current alone (and using fetch? rather than meas:curr?), I can get 35 or so measurements per second, which is better. However, I would like to sample even faster, or, alternatively, have the instrument keep track of peaks/troughs in the current waveform for me and report that information, since it can probably do this at a much higher frequency than it can send me data.

My load right now consists of an ARM microcontroller, which draws about .34 amperes at 5V (which is regulated down on the PCB). The current draw does fluctuate somewhat depending on uC activity, so I thought that perhaps the pulse-current digitization described in the user's manual would be useful for getting large amounts of data very quickly. So, I implemented a slightly modified version of the example code on page 2-15 of the UM (rev. B), in which I use an average current value computed previously as the trigger level rather than 100mA. My aim with this was to have it send me 3600 current values whenever this current threshold was exceeded, which would hopefully be nearly all the time.

However, although a pulse is detected when I use this method, every single value (of the 3600 I receive after read:arr?) is +9.9E37, which is of course nonsense. The commands I am sending the PSU are essentially as follows:

Code: Select all

SENS:RANG 5
VOLT 5
CURR 1.25
OUTP ON
SENS:PCUR:SYNC OFF
SENS:PCUR:AVER 3600
SENS:PCUR:SYNC:TLEV $avg_current
SENS:PCUR:SYNC:DEL 50e-3
SENS:FUNC “PCUR”
READ:ARR?
I think I might be misunderstanding the pulse-current digitization process. Am I going about digitization wrong, or is there a better way to sample data as quickly as possible? It seems to me that the pulse-current examples mostly deal with more square-wave-like loads, whereas mine fluctuates but not quite so dramatically. Thank you for any assistance!

Kenneth P
Keithley Applications
Keithley Applications
Posts: 212
Joined: October 25th, 2010, 1:31 pm
Country: United States
Contact:

Re: 2304A Power Consumption Measurement

Post by Kenneth P » September 6th, 2013, 8:08 am

Hello,

Your findings are consistent with the performance of the 2300 Series instruments. The +9.9E37 is a common value used in the automation indicating that a reading was not available at the time of the request. The pulse current digitization is internally configured to detect a leading or trailing change in current based on a trigger level setting. It will actually take a large number of samples are return in integrated value of the high or low portion of a "detected" pulse or the average (high and low) of a detected pulse. Basically, the 2304 is not the correct instrument for peforming the type of measurement you are attempting. I've supported a similar case. Ultimately, the customer went with a 2636A. In your case, a 2601A would be sufficient. It is capable of sourcing voltage and monitoring current at very high measurement rates. For the 2600B series, the fastest sample rate is 20,000 readings per second with up to 140,000 reading buffer capacity. The value drops to 60,000 with source values and timestamps.

Post Reply

Return to “Series 2300”

Who is online

Users browsing this forum: No registered users and 1 guest