#### KeepItSimpleStupid

Joined Mar 4, 2014
5,090
Isolation: If you do use an isolation transformer, you don;t want isolation in the true sense.
You want N (to instruments) and G (to instruments) of the secondary to be bonded together, and to the incoming ground.

There's shielded line cords. I only used a shielded line cord when driving heaters in an RF (13.56 MHz) chamber. The controllers had RFI filters fitted on the incoming power. I designed it this way so I didn;t have to "fix" anything.

I won't say that a shielded power cord would help.

Could I use a smaller resistor to reduce the noise and still get a definitive spike during discontinuity?
You have a 1% resistor. It could be high by 1/2 an ohm. If it is, you can't deliver 100 mA by definition.
I used 2 ohms as an estimate. 0.5 resistor + 1.5 ohms for the wires and definitely less than 100 mA.
You definitely have to increase the compliance voltage. The power supply can;t make up it's mind between constant current and constant voltage.

Small resistors are sometimes use to isolate capacitive loads. Your problem is that your current limiting can be competing with your required 100 mA. Your resistor may have a 1% tolerance, so it could already be 50.5 ohms. You can't deliver 100 mA into 50.5 ohms and at 5 volts. The 5 Volts wins. I<100 mA.

Does the Transient response of the power supply not matter when the resistor is being used?
Remember two axioms:
1. The voltage across a capacitor cannot change instantaneously
2. The current in an inductor cannot change instantaneously.
3. Everything has parasitics.

Would the noise in the resistor Im seeing be a case of Johnson-Nyquist Noise?
The lower the value of the resistance, the lower the noise, so I don;t think 50 ohms is significant.[/QUOTE]

#### dreimund

Joined Jun 6, 2018
16
Remember two axioms:
1. The voltage across a capacitor cannot change instantaneously
2. The current in an inductor cannot change instantaneously.
3. Everything has parasitics.
I have the setup back to where the 50ohm resistors are in the circuit and changed the voltage limit to 6V. Channel 1 5.02V 100mA(Crimp 50 positions 12" 20AWG Wire) Channel 2 5.07V 100mA(Crimp 50 positions 12" 24AWG Wire) Channel 3 has a max voltage of 5Vdc so that will be something I need to talk to the other about. That channel is essentially useless if we want to maintain the 100mA with a 50ohm resistor.

Small resistors are sometimes use to isolate capacitive loads. Your problem is that your current limiting can be competing with your required 100 mA. Your resistor may have a 1% tolerance, so it could already be 50.5 ohms. You can't deliver 100 mA into 50.5 ohms and at 5 volts. The 5 Volts wins. I<100 mA.
This ^ thank you.

I really do appreciate everybody who contributed and helped me gain more knowledge on this subject.

#### KeepItSimpleStupid

Joined Mar 4, 2014
5,090
Channel 3 has a max voltage of 5Vdc so that will be something I need to talk to the other about. That channel is essentially useless if we want to maintain the 100mA with a 50ohm resistor.
In this case, you have to change the resistor.

Hopefully you do understand that with the values chosen 5 v and 100 mA and 50 ohms when you add tolerances, contact resistances and wiring resistances your trying to occupy the same space at the same time. That just isn't a good idea.

Adding 2 ohms to the chosen resistor, think is a good number. You just need a voltage limit such that 100 mA can be delivered into this load, hopefully with some slop.

I added the resistor tolerance and a guess for circuit resistance. You need to add the other uncertainties to for current and voltage resolution.

#### KeepItSimpleStupid

Joined Mar 4, 2014
5,090
Look at the set voltage as the amount of voltage present on the leads if the circuit was open.

Look at the "test voltage" to have a range depending on the best and worst conditions.
So, 100 mA at ~5 volts is OK. 100.00000 mA @ 5.00000 V is not.
4.5 to 5.5 V at ~100 mA is OK. Leads open is 6V is OK.

#### dreimund

Joined Jun 6, 2018
16
Could I use a 40ohm resistor .100mA roughly 4V? That way there is still some room for the added resistance of the wire and contacts up to 5V?

#### KeepItSimpleStupid

Joined Mar 4, 2014
5,090
using a variation of my wiggle-room formula: I = 4/(40+2) = ~ 120 mA means you should be able to push less current or 100 mA

So, your saying, I want 100 mA and not to exceed 4 Volts assuming your 40 ohm resistor is also 1%.

So, that looks reasonable on the surface.

Now, I would look at how precisely 4 V can be set and how precisely the current can be set in that range.
It still would not help if you were able to set the current in 40 mA steps. The similar argument goes for voltage.

Get the idea?

Joined Jan 15, 2015
6,302
Looking back to your original post and drawing your test configuration looks like this:

So my take on your testing is you are mounting your DUT in a fixture of sorts on a dynamic shaker table. The shaker table is being driven by a large power amplifier and the DUT is subject to random or sine vibration measured using an accelerometer likely with a 10-32 thread connector and now with a low noise cable, so all of that now works.

Your power supply is a Tek Keithly triple output supply having two 0 - 30 volt 3 amp max outputs and one 0 - 5 volt 3 amp max output. In the above drawing if R1 is 50 Ohms and the applied voltage is 5.0 volts the maximum current through the DUT will be 100 mA as long as the supply can provide it. Just for example if you place a shorting bar across one of your supply outputs and adjust your current limit to say 300 mA and remove the short and set your voltage to 5.0 volts when the supply is connected across the 50 Ohm resistor your current will be 100 mA as 5.0 volts across 50 ohms = 0.100 amp. None of this accounts for any lead resistances so in reality as the Rtotal increases and the 5.0 volts remains a constant the current will decrease.What exactly does your test specification call for during the connector (DUT) testing?

Yes, if the maximum Vout of channel 3 of your supply is 5.0 volts with a 50 Ohm load as R1 in your drawing your maximum current will be 100 mA. Reducing R1 will increase the current in t5he circuit as long as your source can provide it, which it should.

The idea in the circuit, as drawn is that in theory the voltage drop across the DUT should be zero volts. In reality it won't be as there is lead resistance and the actual contact resistance of the DUT but it will be close to zero. If for any reason during vibe testing the DUT contact resistance increases the voltage to your monitoring scope will increase to a point where if the DUT were to be open circuit the voltage to the scope vertical input would shoot up to just about 5.0 volts or Vsupply voltage. Remember in theory your DUT resistance should be zero ohms but in reality there will be some small minuscule resistance.

Just as a side note your power supply which has three outputs has three independent outputs which do not share a common. Each output is totally independent of the other two.

Three Fully Independent Outputs
Each output of the Model 2231A-30-3 is fully programmable and can be turned on and off independently
from the other chanqnels; it essentially provides three power supplies in one instrument.
You can power up both the analog circuitry and the digital circuitry of a printed circuit board or a
complete device, all with the same instrument.
All Outputs Are Isolated from Each Other and from Ground
Each output can power a circuit on a different ground reference. The low connection on each output
is not connected to any of the other outputs. That means one supply can test three separate circuits
on three different ground references or can power circuits that are transformer or optically isolated
from each other. Furthermore, all outputs are not referenced to ground, so the Model 2231A-30-3 can
power both floating circuits and grounded circuits.
So in the drawing the only common is the scope vertical input outer shell and if the scope is floated totally isolated you may get some noise. Additionally if you use more than one channel at a time and the scopes channels share common at the inputs your commons are tied there. Don't know if this will matter in your measurement plane or not.

Ron

#### dreimund

Joined Jun 6, 2018
16
"Unless otherwise specified in the referencing document, an electrical load of 100 milliamperes maximum with a detector capable of detecting a discontinuity of 1.0 microsecond or longer"

"Mated connectors shall not be damaged and there shall be no loosening of parts due to vibration. Counterpart connectors shall be retained in engagement and there shall be no interruption of electrical continuity or current flow longer than 1 microsecond"

From Governing Document
"All contacts shall be wired in series with at least 100 milliamperes of current allowed to flow. A suitable instrument shall be employed to monitor the current flow and to indicate any discontinuity or interruption of current flow."

This is the only requirement that we have for the electrical and discontinuity portion. There is no requirement for Voltage (originally I thought there was) but there is a requirement for there to be atleast 100mA of current flow. Thank you

Joined Jan 15, 2015
6,302
OK, so the first quote:
"Unless otherwise specified in the referencing document, an electrical load of 100 milliamperes maximum with a detector capable of detecting a discontinuity of 1.0 microsecond or longer"
They want a load of 100 mA maximum which explains the 50 Ohm resistor with 5 Volts applied. That will afford 100 mA of current through the DUT. Then things go on to mention:
"All contacts shall be wired in series with at least 100 milliamperes of current allowed to flow. A suitable instrument shall be employed to monitor the current flow and to indicate any discontinuity or interruption of current flow."
The concern here for testing is to run 100 mA of current through the DUT while the DUT is subject to a given amount of vibration. The monitoring needs to detect any discontinuity of 1uS duration or longer.
Given the requirements you mentioned I likely would not have gone with the type power supply you mentioned but rather a constant current source for which it is a build or buy. For an example several companies manufacture and market constant current sources with Calex coming to mind. Since voltage is not a primary concern using a constant current source the current can be easily set for each channel (DUT Pin) and a higher voltage becomes easier to detect above the noise levels. Also nice is that with a turn key solution several current sources can be stacked since each is just a edge connector board. The 50 Ohm resistor gets eliminated and the measurement is directly across the connector pins of the DUT. Any discontinuity will show as a fast spike in the voltage drop across the DUT.
So the idea behind the test is while a known rate of vibration is applied to the connector we look at the I*R drop across the connector. Any increase in the R of the connector will be immediately visible as the voltage spikes. Maintaining a constant current if the R of the connector increases and V = I * R the V increase will be apparent. All that is needed for monitoring is something fast enough to catch or capture a 1 uS or greater increase in the R of the DUT pins. As long as no supporting documentation of the entire test cycle duration your scope will do just fine as long as it is setup to catch any small voltage spikes. While your power supplies will work to source current it would not have been my choice but as you already have it and it works I would run with it. Since voltage is not called out I would likely increase R to 120 Ohms and apply 12 volts or so to get my test current.

Ron

#### dreimund

Joined Jun 6, 2018
16
Ron,
Like most of the government specs the main standard says something that contradicts what it references. Mil-DTL-24308 wants at least 100 mA of current flow and tested to EIA-364-28 that states a maximum of 100mA unless otherwise specified in the referencing document. I believe it is to be a minimum of 100mA. I personally did not pick out any of the items we are using for S&V. I came in after the fact and was told to make it work. I programmed all the vibration and shock programs with minimal difficulties. Ran into some noise issues with accelerometers but have spent a majority of the time trying to make sure we would be able to actually capture the 1uS of discontinuity over the noise being picked up. I will play around with increasing the resistance and voltage to see how that works.

This is my last week here and wanted to make this process as easy as possible

Joined Jan 15, 2015
6,302
Ron,
Like most of the government specs the main standard says something that contradicts what it references. Mil-DTL-24308 wants at least 100 mA of current flow and tested to EIA-364-28 that states a maximum of 100mA unless otherwise specified in the referencing document. I believe it is to be a minimum of 100mA. I personally did not pick out any of the items we are using for S&V. I came in after the fact and was told to make it work. I programmed all the vibration and shock programs with minimal difficulties. Ran into some noise issues with accelerometers but have spent a majority of the time trying to make sure we would be able to actually capture the 1uS of discontinuity over the noise being picked up. I will play around with increasing the resistance and voltage to see how that works.

This is my last week here and wanted to make this process as easy as possible
Yeah, I caught that and wanted to comment. The infamous 100 mA. When you have one thing telling you greater that and another telling you not to exceed things get tricky. I also give you credit for your efforts with this being your last week. About 40 years of my career involved testing and collecting data with regard to specifications be they MIL or AMS. I also appreciate that what you have to work with is what you have to work with. I did not review all of your scope specifications but all you need is the ability of having a scope with a fast enough sampling rate to catch a 1.0 uS glitch. Again, your power supply will work, it just is not how I would have gone about it but here nor there. Any set number of pins on the connector? Before I forget, real nice shaker whoever bought it.

I would just continue with what you have going. Also I asked pin count as a single channel of your power supply can test multiple pins. Depending on i9f there is a need to identify which pins fail.

Ron

#### KeepItSimpleStupid

Joined Mar 4, 2014
5,090
I think you just missed the mark on getting 100 mA. because you missed the parasitics and tolerances. 5 V is too close to what it takes to get 100 mA.

Increasing the open circuit voltage does afford you the ability to capture the glitch. So, let;s call the compliance 6V and you need approximately 5V (give or take) to get 100 mA. Your trigger can be just greater than the expected tolerances and parasitics. 1% load resistor + contact resistances + lead resistances.

Measuring that voltage and setting the trigger accordingly might have worked better with more complexity.

Keithley had meters with a "contact test" facility. I can't seem to find info about the operation of it.

EDIT: You may have approached the problem the wrong way. You can easily build an analog feedback ammeter or (Zero Resistance Ammeter) and trigger on current <100 mA. I've built various designs.

Last edited:

#### dreimund

Joined Jun 6, 2018
16
Before I forget, real nice shaker whoever bought it.
not going to lie.... I can not tell if that part is sarcastic or not. Personally, I would of considered a Medium Force system or something a little bigger (300g shock capable). The amplifier came to us defective. Pretty sure it shipped straight from China in a crate. The controller/circuit in the top of the amplifier ultimately had cooling fans that were not turning on and overheating. Had our engineers come down look at the switches/relays inside and determined that one wire was positioned in the wrong spot. Contacted the company to get the okay to move wire and it worked.

The connectors can range from 4 contact positions up to 104 contact positions. Contacts Range from Size 0-22 (80% of the time Size 20 or 22 Contacts, 15% Size 16 or 18 Contacts, Remaining % Size 4 or 0 Contacts). Most common 50 Positions Size 20 contacts (Sample 1 20AWG Sample 2 24 AWG) or 78-104 Positions Size 22 Contacts (Sample 1 22AWG Sample 2 28AWG). In the off chance a discontinuity occurs on non R&D samples it usually is a case of sample prep mistakes. PCB to high on contacts tails pressing against bends of the bottom row, Improperly stripped wire, wrong crimp setting, etc...... Sample Prep is something I have stressed to everybody. I can run 28awg wire with no cable clamps without any issues. You knick one strand on 28awg wire and there is a good chance you will have a pull out or complete break without the cable clamps.

Our Oscilloscopes are Tektronix DPO2002B
MSO2000B and DPO2000B series oscilloscopes can help you verify, debug, and characterize electronic designs. Key
features include:
200 MHz, 100 MHz, and 70 MHz bandwidths
2 channel and 4 channel models
Sample rates up to 1 GS/s on all analog channels
1 M points record length on all channels
5,000 waveforms/second waveform capture rate

I am very grateful for all the advice and information that has been given to me. My knowledge has grown significantly with this project but at the end of the day my degree is in Occupational Safety and Health and (for good or bad) have had to teach myself most of the stuff related to S&V. I know I've used an oscilloscopes before in either physics or computer maintenance but this has definitely been a learning experience for me.

Joined Jan 15, 2015
6,302
not going to lie.... I can not tell if that part is sarcastic or not.
No sarcasm. Just about all the systems I used were MB Dynamics. Not to say those were the best but they were local here in the Greater Cleveland, Ohio suburbs. I did work with a hydraulic system made in Japan about 30 years ago which was a nightmare. I did read a little on the system you have and it looked OK and the specifications were nice for what you seemed to need.

I was just curious as to the connector pins more than anything else and can appreciate that what you have was not your doing. I can also appreciate the fact that Occupational Safety and Health is a good distance from where this project has taken you. As we called it Shake & Bake was a field unto itself. I agree with KISS in that since there is no voltage specification I would increase the voltage as a higher open circuit voltage will be easier to capture on your scope.

Ron

#### KeepItSimpleStupid

Joined Mar 4, 2014
5,090
Externally triggering the scope on a drop in current, I think, might be the best thing to do.

A decent OP-amp with an LT1010 in the feedback loop will work. I used the LT1010 and the obsolete OP-41.

Vos and Ib are the most important specs for the OP-amp.

With 4 channels you might generate the 100 mA reference within the external device and generate a composite trigger.
With a 10 ohm resistor and 100 ohms you get 10V out for the I-V converter. With +-15 V supplies you can still show more than 100 mA.
The circuit is easy, I think, but you may need a bunch of power supplies.

Maybe not... 1/1uS = 1 MHz

#### Siniša Babić

Joined May 15, 2016
3
Hello!
Why don't you put connector on the high side, and detect voltage dropout (low level) as a fault.
Set scope to trigger for low level pulses lasting for 1 microsecond or more...
Also put ferrite beads on power source..

P.S. And yes, don't use wirewound resistor for this, use metal foil. Wirewound one is a transformer coil in disguise....

Regards,

Last edited:

#### rtoth

Joined Oct 5, 2017
4
I am unsure exactly what you are meaning. Without the filters on the Oscilloscope being used the noise can hit close to 4V. The DC power supply sends 5V @ 100ma through a 50ohm resistor. When there is continuity this shows up on the oscilloscope the value is very small. When discontinuity occurs with the resistor in place the oscilloscope will pick up the voltage spiking up.

Accelerometer Noise can be seen here. This was resolved with the low noise cables.View attachment 153788

This would be the settings we use to monitor for discontinuity. Noise filter on oscilloscope has to be used see below
View attachment 153789

View attachment 153792
View attachment 153790
Your weak-links are the connectors, specially if the vibrations are high. Many times when I need to take a low noise measurement, I use shield not as a current carrying conductor, just a shield and I will run my ground from the chassis, bypassing connectors. If you use this method, you must run this ground parallel to you instrument cable.

Last edited by a moderator:

#### Tank Circuit

Joined Mar 5, 2017
30
I have attempted to locate cracked solder connections using my scope probes to monitor certain test points while tapping on the circuit board. It worked countless time, but then there the times where it led me astray. In those instances, I was actually causing the glitches as I tapped, internal to the scope probes and the contacts to the board.
Be careful that you are not seeing noise as a result of your test bed and associated wiring and probes.