EPIC 20.000 (imminent release) has semi-automatic and automatic (as well as manual) signature comparisons, board picture and test-tree functions. So more like Huntron Workstation than FADOS.
Because of the pulsed mode nature of testing power semiconductors we sample all ADCs (4x 24bit) simultaneously. I don’t know if the Zero can do simultaneous sampling, probably not important if you’re interfacing to a non-pulsed CT.
IMO This is the most important reason to test semiconductors.... for making models used in CAD. The other, IMO, is for matching/sorting. Especially FETs.
I have a Chinese made curve tracer (CRT wow) which allows me to see two device characteristics displayed at the same time on the screen. Great for matching.
THx-RNMarsh
Transistor Curve Tracer
Last edited:
@jackinnj, is this the one? LINK
Then along with the opto Isolateors, (which ones do you use?)
How do you interface this for measuring?
Don't you have to write your own code also?
Is there an open source that you use?
I do have an earlier version of that Arduino I think.
let me find it.
Cheers,
I have a bunch of Arduinos -- the one I use most often is the Linduino which Analog Devices (now owns LLTC) sells. I have used the HCNR200 analog isolator -- these are original HP, before they sold the opto biz to Avago, also use IL300 which are now Vishay. (Mine are Siemens so also quite old!) I used the opto's as I didn't want to fry my laptop when working with tubes and their high voltages. Nowadays I just buy refurbished from micro-Center for a couple hundred bucks and get last year's best model.
ADI has development boards for the Linduino -- DACs ADCs thermal loggers etc.
I noticed the CTL503 is now availabe even if the manual is still very much WIP:
https://electron.plus/wp-content/USERMANUAL/CTL503_User_Manual.pdf
There no not seem to be any user reports on diyaudio. How mature is the product?
https://electron.plus/wp-content/USERMANUAL/CTL503_User_Manual.pdf
There no not seem to be any user reports on diyaudio. How mature is the product?
Hi capslock,
The hardware and firmware has been complete for sometime, and we have built a number of units. The PC software still has some functions incomplete, and there's a wish-list of features to add over time.
Manuals: yes, WIP indeed.
We started out with the CTL502 (pictured on the w/s as well) - we built a batch, but deceided not to release it in the end as it was a raw PCB type product and we wanted to conserve our stock microcontrollers - Hence the CTL503 (24bit converters, 100V/3A, ext. PSU, relays not switches and a pulsed collector function as well as pulsed base/gate).
We aren't pushing the CTL503 until the manual is complete and we've edited some videos and application notes. There are some better pix on our US distributors site, and they have some stock - we also have stock here in the UK.
Electron Plus CTL503 Curve Tracer ($232.00) : Saelig Online Store
The hardware and firmware has been complete for sometime, and we have built a number of units. The PC software still has some functions incomplete, and there's a wish-list of features to add over time.
Manuals: yes, WIP indeed.
We started out with the CTL502 (pictured on the w/s as well) - we built a batch, but deceided not to release it in the end as it was a raw PCB type product and we wanted to conserve our stock microcontrollers - Hence the CTL503 (24bit converters, 100V/3A, ext. PSU, relays not switches and a pulsed collector function as well as pulsed base/gate).
We aren't pushing the CTL503 until the manual is complete and we've edited some videos and application notes. There are some better pix on our US distributors site, and they have some stock - we also have stock here in the UK.
Electron Plus CTL503 Curve Tracer ($232.00) : Saelig Online Store
Last edited:
I was not aware of the CTL503/502 development -- congratulations for doing this!
Can you shine some light on how the base current is determined with BJTs? The website says there are four selectable base resistors. Does the curve tracer measure the voltage drop across this resistor directly to determine the true base current? Or does it calculate the voltage drop across the resistor from the applied voltage and subtracting the (approximate) Vbe voltage of the BJT?
Can you shine some light on how the base current is determined with BJTs? The website says there are four selectable base resistors. Does the curve tracer measure the voltage drop across this resistor directly to determine the true base current? Or does it calculate the voltage drop across the resistor from the applied voltage and subtracting the (approximate) Vbe voltage of the BJT?
Base current sourced by an active current generator circuit, so is not affected by changes in Vbe. The current generator is driven by a DAC, and has 4 range setting resistors, it is calibrated at the factory and can also be calibrated in the field (each CTL503 is supplied with 4 known resistors, so can be calibrated using a half-decent DMM).
There is an special clamp circuit that is engaged if the base is run in pulse mode (the output compliance voltage can rise to approx. +/-15V if it is trying to sink/source zero current).
In voltage (FET) mode, the output gate drive resistor is set to 270R.
Vbe is monitored by the ADC in the microcontroller (12 bit) over the +/-1.5V range - this is to help Vbe matching.
There is an special clamp circuit that is engaged if the base is run in pulse mode (the output compliance voltage can rise to approx. +/-15V if it is trying to sink/source zero current).
In voltage (FET) mode, the output gate drive resistor is set to 270R.
Vbe is monitored by the ADC in the microcontroller (12 bit) over the +/-1.5V range - this is to help Vbe matching.
Last edited:
kannan_s:
We will publish accuracy specs soon, these are slightly more complex to define as the accuracy in pulsed mode is a quite difficult to measure directly.
Whilst it's easy to dump a few amps for 100us into a non-inductive resistor and measure pulse height using an oscilloscope, doing this at high enough resolution is beyond any oscilloscope I've come across (most of ours are 8 or 10bit, even a 12 bit scope wouldn't have enough accuracy to be meaningful).
Development of the calibration system for the CTL503 took a huge amount of time, and as mentioned earlier, it is carried out with a DMM and 4 resistors. It takes 10-20minutes at least to do, so we developed an automatic system and fixture to speed it up (closed box calibration). Here's a screen shot of the CTL503 calibration page in EPIC:
Ultimatley the accuracy of the CTL is largely defined by the temp-coefficient of internal resistors, and as you will appreciate producing an affordable unit does require some compromise. Still - in our experience - it is far more accurate than any legacy curve-tracer (short of a Keithley 4200).
mbrennwa:
Documentation is coming and we will detail the protocol as well, we have plans for an API at some point also.
Like most of our products, the heavy-lifting is done in the PC, so there is a fair complexity to interfacing to the unit - including some safe-guards regarding the HV PSU. We spent alot of time trying to make it 'bomb-proof', although time will tell how this works out.
We will publish accuracy specs soon, these are slightly more complex to define as the accuracy in pulsed mode is a quite difficult to measure directly.
Whilst it's easy to dump a few amps for 100us into a non-inductive resistor and measure pulse height using an oscilloscope, doing this at high enough resolution is beyond any oscilloscope I've come across (most of ours are 8 or 10bit, even a 12 bit scope wouldn't have enough accuracy to be meaningful).
Development of the calibration system for the CTL503 took a huge amount of time, and as mentioned earlier, it is carried out with a DMM and 4 resistors. It takes 10-20minutes at least to do, so we developed an automatic system and fixture to speed it up (closed box calibration). Here's a screen shot of the CTL503 calibration page in EPIC:
Ultimatley the accuracy of the CTL is largely defined by the temp-coefficient of internal resistors, and as you will appreciate producing an affordable unit does require some compromise. Still - in our experience - it is far more accurate than any legacy curve-tracer (short of a Keithley 4200).
mbrennwa:
Documentation is coming and we will detail the protocol as well, we have plans for an API at some point also.
Like most of our products, the heavy-lifting is done in the PC, so there is a fair complexity to interfacing to the unit - including some safe-guards regarding the HV PSU. We spent alot of time trying to make it 'bomb-proof', although time will tell how this works out.
