A cheap impedance meter for plating nicrome tetrodes

[Jumpei Matsumoto]

Hi all,

Does anyone know a cheap and nice impedance meter to test nichrome wire tetrodes?

Or is anyone using a LCR meter for general purpose (e.g., http://www.keysight.com/en/pd-1984403-pn-U1731C/) instead of a impedance meter designed for electrophysiology (e.g., http://www.bakelectronicsinc.com/IMP-2A%20&%20IMP-2AMC.pdf)?

I am going to setup a classic gold plating system (Fig. 1 of Nguyen+ 2009, https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2794082/), because nanoZ (https://neuralynx.com/hardware/nanoz-kit) is expensive (around 5,000 dollars in Japan). But I also found that a impedance meter designed for electrophysiology is also expensive (> 2,000 dollars)...

The main difference between the one for general purpose and the one for electrophysiology seems to be the current amplitude (general: 0.74 Vrms, so 7.4uA with 100kohm; electrophysiology: 5 - 100 nA). Can this relatively large AC current (~10uA) damage a wire tetrode and/or distrub gold plating?

Any idea for setting up the cheap gold plating system is very helpful!

Best, 

Jumpei

[Jon Newman]

1. Plating -- you need a simple biploar DC current source with a relatively wide compliance voltage. There are two topologies I have used and work well. 

- BJT totem pole (used by intan)

- R-sense across instrumentation amp (Im using this).

- I've attached figures of both.

2. Impedance measurement -- I just took apart a nanoz today, there is nothing going on in there. So far I see

- A multiplying DAC and a bunch of muxing. Have not sussed out the rest of the analog stuff, but looks simple [ need figure out]

- Switched cap inverting/boost regulators to make +/-5V [replace with LT uModule, e.g. LT3463]

- MCU [replace with Teensy]

- FTDI chip for USB comm. [ Not needed with teensy]

If you are interested in making a free and open source version, start a repo and I'm happy to help and send pull requests.

Jon

--
You received this message because you are subscribed to the Google Groups "Open Ephys" group.
To unsubscribe from this group and stop receiving emails from it, send an email to open-ephys+unsubscribe@googlegroups.com.
To post to this group, send email to open-...@googlegroups.com.
Visit this group at https://groups.google.com/group/open-ephys.
To view this discussion on the web visit https://groups.google.com/d/msgid/open-ephys/6e9eb9b4-86da-45c7-91f5-b4928700273f%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

Jonathan Newman

Postdoctoral Fellow, MIT

http://www.mit.edu/~jpnewman/

[Jumpei Matsumoto]

Hi, Jon.

Thank you very much for the detailed circuit for DC current generation and kind offer to help developing an open source system.

For measuring impedance, today I found "auto balancing bridge method", which is really simple to implement.

Although I am interested in making "nanoZ like" automatic system combining those circuit and share it, eventually, I also need something to quickly start with. In this point of view, do you think a LCR meter for general purpose can be a substitute for a impedance meter designed for electrophysiology (e.g., http://www.bakelectronicsinc.com/IMP-2A%20&%20IMP-2AMC.pdf)?

Thanks,
Jumpei

2017年8月30日水曜日 10時42分16秒 UTC+9 Jon Newman:

1. Plating -- you need a simple biploar DC current source with a relatively wide compliance voltage. There are two topologies I have used and work well. 

- BJT totem pole (used by intan)

- R-sense across instrumentation amp (Im using this).

- I've attached figures of both.

2. Impedance measurement -- I just took apart a nanoz today, there is nothing going on in there. So far I see

- A multiplying DAC and a bunch of muxing. Have not sussed out the rest of the analog stuff, but looks simple [ need figure out]

- Switched cap inverting/boost regulators to make +/-5V [replace with LT uModule, e.g. LT3463]

- MCU [replace with Teensy]

- FTDI chip for USB comm. [ Not needed with teensy]

If you are interested in making a free and open source version, start a repo and I'm happy to help and send pull requests.

Jon

On Tue, Aug 29, 2017 at 7:46 PM, Jumpei Matsumoto <mtj...@gmail.com> wrote:

Hi all,

Does anyone know a cheap and nice impedance meter to test nichrome wire tetrodes?

Or is anyone using a LCR meter for general purpose (e.g., http://www.keysight.com/en/pd-1984403-pn-U1731C/) instead of a impedance meter designed for electrophysiology (e.g., http://www.bakelectronicsinc.com/IMP-2A%20&%20IMP-2AMC.pdf)?

I am going to setup a classic gold plating system (Fig. 1 of Nguyen+ 2009, https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2794082/), because nanoZ (https://neuralynx.com/hardware/nanoz-kit) is expensive (around 5,000 dollars in Japan). But I also found that a impedance meter designed for electrophysiology is also expensive (> 2,000 dollars)...

The main difference between the one for general purpose and the one for electrophysiology seems to be the current amplitude (general: 0.74 Vrms, so 7.4uA with 100kohm; electrophysiology: 5 - 100 nA). Can this relatively large AC current (~10uA) damage a wire tetrode and/or distrub gold plating?

Any idea for setting up the cheap gold plating system is very helpful!

Best, 

Jumpei

--
You received this message because you are subscribed to the Google Groups "Open Ephys" group.

To unsubscribe from this group and stop receiving emails from it, send an email to open-ephys+...@googlegroups.com.

To post to this group, send email to open-...@googlegroups.com.
Visit this group at https://groups.google.com/group/open-ephys.
To view this discussion on the web visit https://groups.google.com/d/msgid/open-ephys/6e9eb9b4-86da-45c7-91f5-b4928700273f%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Jumpei Matsumoto]

Hi Jon,

I finally made a repo of the circuit!

The present one is not automatic nor multiplexing. It even does not have an impedance display. 

However, I am almost happy with it, because it works quick (it needs only <10 sec/ch to electroplate).

Thank you very much!

Jumpei

2017年8月30日水曜日 22時48分38秒 UTC+9 Jumpei Matsumoto:

[Jon Newman]

Awesome!!! I feel like adding muxing will be quite easy now that the hard part is done. Thanks for putting so much effort into this!

To unsubscribe from this group and stop receiving emails from it, send an email to open-ephys+unsubscribe@googlegroups.com.

To post to this group, send email to open-...@googlegroups.com.
Visit this group at https://groups.google.com/group/open-ephys.

To view this discussion on the web visit https://groups.google.com/d/msgid/open-ephys/ab48da21-0f31-41ed-80e2-403398d05029%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

[Jumpei Matsumoto]

Thank you too, for your kind help!

I enjoyed the development. But unfortunately now I don't have much time for further development. 

I hope that someone in the community could continue/help development for adding muxing :) 

2018年5月8日火曜日 1時25分58秒 UTC+9 Jon Newman:

To view this discussion on the web visit https://groups.google.com/d/msgid/open-ephys/ab48da21-0f31-41ed-80e2-403398d05029%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

[Christian Skoven]

Dear Jumpei,

So interesting to read your progress.!

I was talking to Jon about the same issue a while back - but was a bit overwhelmed before I got started.

I'm thinking of / planning to build your Tetroplater - if that works for you :)

Can you elaborate a bit more on the "electrode holder" and how you switch between the electrodes you're plating.?

I will be using 32ch Omnetics-connectors.

Br,

Christian Skoven

[Jumpei Matsumoto]

Hi Christian,

The device works well for me. I am happy if it could also help you. 

My "electrode holder" (pic1.jpg) consists of a magnetic base (Kanetec MB-B), a manipulator (Narishige; not necessary to be so precise), and a custom made adapter board. 

The adapter board is similar to this, but made for a Samtec connector and our manipulator. The board makes switching between electrodes easy, as I can connect the probe needle (22G) with an electrode by touching the corresponding hole on the board with the needle (pic2.jpg). The eagle file of my adapter board is attached. You can customize it by replacing the foot print for Samtec connector with the one for the Omnetics 32ch. The foot print of the Omnetics can be copied from the open-ephys headstage eagle files.

Please feel free to ask me more information.

2018年6月4日月曜日 21時54分30秒 UTC+9 Christian Skoven:

[ONECore]

Well done, Jumpei! I am wrapping my head around this and will look into a multiplexing system (and a 3D printed enclosure). I have the parts list (current set up) in my shopping carts, but had an idea: Could me move the dual power supply to the board? The dual power supply is $13.70USD. It could be cheaper ($3.98) (and a more compact system!!) using something like the rail slitter similar to the Dynamic Clamp System: https://dynamicclamp.com/methods/

New parts:

32			296-6550-5-ND	IC VREF GND REF ADJ 8DIP		2 Immediate	0	2.30000	$4.60
33			399-4389-ND	CAP CER 1UF 50V X7R RADIAL		3 Immediate	0	0.47000	$1.41
34			CP-031D-ND	CONN PWR JACK 1.3X3.5MM SOLDER		10 Immediate	0	0.57000	$5.70
35			P5193-ND	CAP ALUM 100UF 20% 63V RADIAL		10 Immediate	0	0.24800	$2.48

And maybe one more AE9986-ND? And don't forget the power supply: https://www.amazon.com/gp/product/B0002E3FBK/ref=oh_aui_detailpage_o03_s00?ie=UTF8&psc=1

Thoughts?

Could you explain the item you have listed as 6, the resistor calibration? 

Thanks!
Andrew

[Jumpei Matsumoto]

Hi Andrew,

Sounds cool! It is really nice if you will expand the system in that way. 

- As for the power supply, I have not tested different power supplies (I only tested the commercial dual power supply module and a pair of 9V batteries). There are a few concerns need to be tested: 1) The sin wave generator in the circuit is very sensitive. For example, slight change of a resistance value in the circuit could affect the stability of oscillation (this is why I used the relatively precise resistance, RNF14FTD4K53CT-ND, for R3). If the power supply is not so stable, it also might affect the oscillation; 2) If electric noise is emit from the power supply, it may disturb the measurement. But, anyway because I have not tested, so I have no idea if these concerns are actually significant or not. I would suggest you to just try :)

- In my digikey list, No. 6 seems to be a capacitor. Could you kindly provide the product ID?

Best,

Jumpei

2018年6月30日土曜日 0時44分51秒 UTC+9 ONECore:

[ONECore]

I updated the Gerber files to allow for the power supply rail system to be directly on the board. I sent you a pull request.

I have the parts on order. Will update when assembled.

I meant number 6 on the picture (an example of setup (with numbers).jpg). You call it the resistor calibration. How did you implement that?

Thanks,

Andrew

[Jon Newman]

I'm going to put in my two cents on this design since it seems like you two might have some momentum right now:

- Make it compatible with NanoZ adapters (use the same connector). This will make it immediately useful to hundreds of labs

- Obviously use very low test voltage/currents

- Simple is better. e.g. just having 1 kHz is fine. 

- Build around teensy.

- Jon

To unsubscribe from this group and stop receiving emails from it, send an email to open-ephys+unsubscribe@googlegroups.com.

To post to this group, send email to open-...@googlegroups.com.
Visit this group at https://groups.google.com/group/open-ephys.

To view this discussion on the web visit https://groups.google.com/d/msgid/open-ephys/ff94ec8e-3430-475c-b8a4-d9d5920fe16f%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

[Jumpei Matsumoto]

Dear Jon,

Thank you very much for the good ideas. As for the test signal voltage, I designed the circuit for ~2mV p-p. Today I tried to reduce it (just by turning the potentiometer for stronger attenuation). If I use a Faraday cage (just aluminium foil, "shielding.jpg"), I could reduce the test voltage to ~0.4 mV p-p (see the result in "low voltage test signal 180703.jpg"). With 0.4mV, the current will be 2 nA for 200 kohm electrode. Do you think it is enough small? Although wave is not very clean, it would be ok, if Andrew will use a spectrum analyzer for amplitude readout or with further analog/digital filtering. 

Dear Andrew, 

Thank you very much for sharing the nice design. 

I would be grateful if you could tell me the test result and give me a picture of the assembled circuit, then I would like to update the repo with a revised manual.

I attached a large picture of the component 6 ("resistors for calibration.jpg"). It is just a breadboard + resistors (to calibrate the board).

In addition to Jon's comment, I would suggest one thing about further development for multiplexing:

- it might be better to replace most of the parts with the corresponding SMD. Then you can ask a company to (get and) assemble those parts (not so expensive, maybe) to save space in the board, save time for assemble and avoid error during assembling. Soldering  >50 parts manually may sound not so nice...

I hope these could help. 

Best regards,

Jumpei

2018年7月3日火曜日 6時21分25秒 UTC+9 Jon Newman:

To view this discussion on the web visit https://groups.google.com/d/msgid/open-ephys/ff94ec8e-3430-475c-b8a4-d9d5920fe16f%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

[ONECore]

Hi Jon and Jumpei,

I have built a system (with the power supply onboard), but wanted to get to the muxing and teensy system. 

It's not at all complete, but I did test the audio system, and I can output a very good VOut  at 30 mV, 10000Hz (no bandpass required)(not tested, but presumably it can read these voltages very well also). In only 300 lines of code, I have a skeleton of code that should be a way to calibrate the output (DC and waveform), has a variable number of channels, tests the resistance, and does the plating to a desired resistance for each channel (seems like some people might like to experiment with different resistances). It will require two muxing systems that I am 'stealing' from Sparkfun (https://learn.sparkfun.com/tutorials/multiplexer-breakout-hookup-guide?_ga=2.99037278.1356289842.1532393743-473632450.1532393743#74hc4051-breakout-overview), but with 16 channels. Unfortunately, I cannot find a good alternative that is through hole, so these two components will require surface mount soldering. I will update the eagle files to remove the components I no longer require, and hold the teensy. Not sure if I should design the Eagle files for more channels? How many do you use? And what connectors? It seems like NanoZ connectors are stupidly proprietary? How can I connect to them on the cheap? 

The project is far from complete, and no doubt full of tons of errors though out(! Like, I am not going to hit verify until I do a lot more work!), but I just wanted to say I haven't abandoned ship.

Current code (not yet commented out, so it might not make sense, don't look at detail, just big picture look at the individual functions) can be found: https://gitlab.com/OptogeneticsandNeuralEngineeringCore/TeensyPlater/blob/master/TeensyPlater

PS: the DC current system is great. I will be using that to make a lesion maker as well. Check out my gitlab for a cry for help, that I will be taking on myself now. https://optogeneticsandneuralengineeringcore.gitlab.io/ONECoreSite/projects/CallForHelp/

Andrew

ONE Core

[ONECore]

PPSS, the shielding will be well done with a 3D FDM printed enclosure, which I will design. You can the spray the outside of the case with a conductive paint ($30 for a bottle which should last...like 1000 of these). It is very conductive. I was testing it with a dynamic clamp system and accidentally shorted it (I fried the teensy, several op amps, and lord knows what else) hence why this project got so delayed). Now I understand that I should put this paint on the outside of the case. Should be a sweet system when done.

Andrew

[Jon Newman]

Cool.

Some suggestions:

- Use kicad if possible. I'm trying to switch myself. The newest version looks amazing. Would be great to work on a project that used it. 

- The nanoz connectors are not proprietary. They are from the MOLC series from Samtec

https://www.samtec.com/products/molc

They are an excellent choice IMO because they are mechanically sound and withstand a ton of cycling. I would highly recommend retaining this connector and the nanoz pinout so people can use existing EIB adapter boards. 

- You are dealing with measuring pretty large resistances and pretty small capacitances. Therefore routing and muxing choices are going to matter a lot for getting accurate readings. Be careful about what you choose here Look for muxes with very low charge injection and parasitic capacitance. These are typically not going to be through hole. Through hole components introduce lots of parasitic in the leads, which you don't want have for this design. Here is an example of a muxing system I made for the nanoz that is fairly high performance with the exception of the overly long traces over ground.

https://github.com/jonnew/open-ephys-pcie/tree/master/nanoz-adapter-128-256

https://github.com/jonnew/open-ephys-pcie/blob/master/nanoz-adapter-128-256/datasheets/dg4051e_analog-mux.pdf

- I'm not a big fan of 3D printing for electronics enclosures. Extruded aluminum enclosures are better performing and often cheaper.  Additionally, projects that are needlessly lightweight are very annoying for lab use because the board tends to be mechanically dragged around by the cabling. This is probably a bad thing for a delicate plating application. Spray on conductive coating inside of the 3D printed enclosure is OK but will never match the shielding provided by a 2 mm aluminum. PCB to shield contact can be made with a spring contact.

That said, if you want to maintain the nanoz's mechanical design and structure, 3D printing might be the best option because it requires the connector hole to be milled into the face of the enclosure. If you want to use the side panels for the connector instead, then panels can be made from PCB material V-CUT or tabbed on the main PCB. Since these will have a ground plane, you will get really good shielding for free. e.g. like was done on the cyclops

- For sending a receiving data to/from the teensy, consider using a ready made message serialization library instead of ad-hoc characters. ArduinoJSON is particular impressive.

- Tell  use once you have a complete schematic and we can do a hardware review.

- Jon

To unsubscribe from this group and stop receiving emails from it, send an email to open-ephys+unsubscribe@googlegroups.com.

To post to this group, send email to open-...@googlegroups.com.
Visit this group at https://groups.google.com/group/open-ephys.

To view this discussion on the web visit https://groups.google.com/d/msgid/open-ephys/4e9e11f7-0518-454f-957a-0f2625b0f344%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

[Jumpei Matsumoto]

Hi Andrew,

Sorry for late reply.

I don't fully understand the code but I have a small comments on the code:

- Because 1kHz sine signal is common to test the electrode impedance, the parameter of sine1.frequency() should be 1000 not 10000?

- Also you may need to set the frequency in AudioAnalyzeToneDetect class objects.

I will also be happy to test when you will finish making something working.

Best,

Jumpei

2018年8月14日火曜日 12時38分17秒 UTC+9 ONECore:

[Andrew Scallon]

Hello everyone.

See here for some files: https://gitlab.com/OptogeneticsandNeuralEngineeringCore/TeensyPlater/blob/master/TeensyPlaterV.5.zip

I first created a flow diagram to organize my thoughts. It may be useful to you, or complete gibberish.

I then learned enough Eagle to smash together a schematic and board (am I doing something wrong? it feels unnecessary to define a pin like  5 times...). I also wrote up some Arduino code to help control this thing.

Anyhow, I tested the 30 mV sinusoid output from the Teensy Audio shield, and that is great. If it can output that, presumably it can read it (there's a nice function to average over 10 cycles peak to peak, when called). I used the same constant current generator and auto balancing bridge as before (with added two Jumpers to allow for the DC current to flow in either direction (someone here mentioned it might be cool to lessen the resistance of an electrode sometimes). I am using an Omnetics A79042 connector. I chose the ADG1206yruz Mux. It is surface mount and should be fun to test your soldering skills on (use solder paste and pray). They are so small, they can fit under the Teensy. A step up supply (3.3 to 5) was necessary to drive the constant current circuit (using 3.3V to drive this circuit brought up the current to 3.5 uA. Not terrible, but it seemed inconsistent). The audio board will sit on top of the Teensy.

The code passes verify (can I ever not forget a semicolon??). It is pretty well commented up so you might have a chance of understanding it. 

Let me know if you can find the time to check this out and give feedback on this before I order these parts up. Any feedback is great, as this is my first time through this kind of thing.

Andrew

You received this message because you are subscribed to a topic in the Google Groups "Open Ephys" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/open-ephys/WM3JAaYDXw8/unsubscribe.
To unsubscribe from this group and all its topics, send an email to open-ephys+...@googlegroups.com.

To post to this group, send email to open-...@googlegroups.com.
Visit this group at https://groups.google.com/group/open-ephys.

To view this discussion on the web visit https://groups.google.com/d/msgid/open-ephys/fdeae3b2-e59c-4554-aa94-cb7426720e75%40googlegroups.com.

[Jumpei Matsumoto]

Hi Andrew,

Thank you for your work and sharing it.

I think that you would need an attenuator and and an amplifier, before and after the auto-balancing bridge, respectively.

The input to the electrode should be as small as "3-4 mV" to make the test signal in the physiological range and to prevent possible plating caused by the test signal. (Note that what you see in an oscilloscope with the original Tetroplater are x10 amplified for display)

Or dose Teensy  have enough resolution for directly outputting and inputting 3-4 mV sin wave?

Personally I think that 64 ch (or more) multiplex sounds much more useful than 16 ch as many people use tetrodes with a hyperdrive for large scale recording. For example, is it difficult to make the 64 ch with two 8-ch multiplexer (8x8=64)? Why do you need two 16-ch multiplexer (ADG1206) for the 16 ch Omnetics electrode connector?

Best regards,

Jumpei

[Jonathan Newman]

Jumpie, is it possible to generate a DOI for this repo:

https://github.com/MatsumotoJ/Tetroplater

I would like to cite it a paper. You can use this: http://about.zenodo.org/.

[Jumpei Matsumoto]

Hi Jon,

Thank you. I made the DOI:

https://doi.org/10.5281/zenodo.3334859

2019年7月15日月曜日 1時12分41秒 UTC+9 Jonathan Newman:

[Cecilia Martinez]

HI Jumpei! Thanks for sharing your tetroplater! I'm trying to build one. Could you tell me the power supply specs? The Amazon link is broken. I was wondering about the current needed for plating. 

Thanks,

Cecilia

[mtj...@gmail.com]

Hi Cecilia, sorry for the broken link. The current needed for plating should be very small, so you may not need very special power supply. The power supply specification should be a dual power supply, whose voltage is +-9 to +-12 V. I suggested some options in a github issue, which may be useful:

https://github.com/MatsumotoJ/Tetroplater/issues/2#issuecomment-509847469

Best regards,
Jumpei

2024年10月4日金曜日 21:37:21 UTC+9 Cecilia Martinez: