BNC as low cost alternative to high quality 4mm banana post ? for resistance box

[Overspeed]

Hello

I have a batch of high quality BNC connectors both chassis and wire connectors . question : does some one have already use BNC for DC resistance box ? Xdevs have been built with Triaxial connector a calibration box so ?

I plan to lathe turn rear collar in PA6 or Delrin to match the cable diameter and get a tight fitting

Alex Nikitin on this forum have used DB connector perhaps BNC can work too ?

Regards
OS

[Alex Nikitin]

I use BNC connectors for my 7D resistance reference box, from 10K upwards it is a good solution, providing the connectors are good quality (machined body, Beryllium Bronze contacts, PTFE isolation) and not cheap Chinese fakes from Amazon  .

Cheers

Alex

[Overspeed]

I use BNC connectors for my 7D resistance reference box, from 10K upwards it is a good solution, providing the connectors are good quality (machined body, Beryllium Bronze contacts, PTFE isolation) and not cheap Chinese fakes from Amazon  .

Cheers

Alex

Hello

Thanks I agree on the look a like chinese connectors , I plan to use Radiall as this one with the matching socket as I have a batch of them , I have no budget for Pomona 4mm

Regards
OS

[Alex Nikitin]

Radiall is good, probably my favorite, Amphenol is also good. However both do make BNC sockets with either PTFE or Nylon isolation, PTFE is much better, so it pays to check carefully the part number.

Cheers

Alex

[EC8010]

The advantage of BNCs for resistance boxes is that it gives you a screened connection, and that allows you to connect a resistance box in place of a resistor within an active circuit and not add hum. I have converted my resistance boxes to BNC. I fitted a BNC for each terminal, so that means I can cascade resistance boxes (100M box in series with 100k box) and terminate the end of the line with a BNC shorting cap or electrometer. It has worked out really well. The other thing to do is to ensure that your resistance boxes are fully screened; I lined my wooden boxes with earthed aluminium foil.

[Overspeed]

Radiall is good, probably my favorite, Amphenol is also good. However both do make BNC sockets with either PTFE or Nylon isolation, PTFE is much better, so it pays to check carefully the part number.

Cheers

Alex

Yes for PTFE , I have read that 75 Ohms have PTFE and most of the 50 Ohms have nylon ( PA66 ) but that possible to locate BNC 50 Ohms with PTFE , I will check in my available batches

The only concern I have seen on BNC is the maximum possible matting before contact surface damage as from main brand as Amphenol they rate at 500 connections before possible contact damage ( galling / fretting)

Regards
OS

[Overspeed]

The advantage of BNCs for resistance boxes is that it gives you a screened connection, and that allows you to connect a resistance box in place of a resistor within an active circuit and not add hum. I have converted my resistance boxes to BNC. I fitted a BNC for each terminal, so that means I can cascade resistance boxes (100M box in series with 100k box) and terminate the end of the line with a BNC shorting cap or electrometer. It has worked out really well. The other thing to do is to ensure that your resistance boxes are fully screened; I lined my wooden boxes with earthed aluminium foil.

Hello

Yes for screened connection , that the triaxial '' of the poor '' compare to regular basic 4mm banana even is I prefer old style spade / fork .

An interesting subject is does there is some possible improvements with a guard system / net setup ?

Regards
OS

[coppercone2]

I don't think it makes sense unless its high value enough or low current

I use my boxes for all kinds of prototyping and repair work and it would be annoying to have a BNC on it. All my 10K boxes (commercial) have banana connectors with a guard connection on it.


For practical work, you want the binding post to attach a simple wire



I think if you start getting into the mega ohms it makes sense but I still suspect you will get decent performance without a coaxial connection


And usually their not designed for AC, but if its a box designed for AC at high resistance it might be way more applicable


If your stimulus current is very low, then it makes more sense, regardless of the value however. So I think its application dependent! A multimeter uses a reasonable current. Science might not, then you get into BNC

If you do low current AC measurement then I definately want a coaxial box, I see what it does on a miliohm meter (hP), noisy as hell.

so if you plan on using it with a source meter or adjustable current ohm meter, you might find more use with it.



From a mechanical prospective, the BNC is easy to damage, and the interconnects/adapters (for typical use with non BNC cables) are flimsy, and the BNC cable itself is weak and inflexible. For testing circuits with the box for feedback, I think you want short wires more then anything, aka place the box leaning against the circuit with a insulator and 2 short wires attached to the binding posts that go into the PCB. I think having a coaxial cable there to try to get 'range' from it will have bad results compared to having the box as close as humanely possible.



Unless you build your box with 4x coaxial connector for direct hookup to a impedance analyzer, but why? Designing a box to hook up to a 4x bnc impedance analyzer... seems useless unless its just a special experiment to characterize a box that is only existing to be characterized. Having 4x banana (kelvin) still lets you hook up to the IA with an adapter.




What do you want to use the box for?





But anyway, the penalty for using a Dual banana to BNC coaxial cable (if required), or a dual banana to dual banana coaxial cable, is very small compared to a BNC connector. Then your box can be used for general purpose or calibration with the right cable, so long you maintain the correct connector spacing.

[Overspeed]

Hello

Following Amphenol specs a BNC can handle 360 Watt of power but that AC current ( in regular use ) even with 50 percent ratio that still 180 Watt , not bad for measurement

I use 15 V and 0.2 Amp on resistivity system

BNC are cheap from surplus and available in high quality brands , Pomona 4 mm are just $$$ for a individual DIY or even a small company , I can rebuilt and fit new one when I assess problems 

[TimFox]

I assume the Amphenol power spec relates to the rms voltage into a matched 50 ohm load.
That would be 134 V rms = 190 V pk, which I'm sure is adequate for your purpose.
Historically, there were other dielectrics used besides Teflon and Nylon, but you want Teflon for this purpose.
Nylon insulation suffers from water absorption in normal atmospheres.
It's easy to find BNC male cable and female panel connectors with gold-plated center contacts, although the shield connection is not gold-plated.

[Overspeed]

I assume the Amphenol power spec relates to the rms voltage into a matched 50 ohm load.
That would be 134 V rms = 190 V pk, which I'm sure is adequate for your purpose.
Historically, there were other dielectrics used besides Teflon and Nylon, but you want Teflon for this purpose.
Nylon insulation suffers from water absorption in normal atmospheres.
It's easy to find BNC male cable and female panel connectors with gold-plated center contacts, although the shield connection is not gold-plated.

Hello

Yes PA66 is not humidity insensitive and that can be a problem

Strangely Pomona don t provide matting number is their 4 mm specs even on the top of the range model .

Regards
OS

[coppercone2]

the dual posts are kV rated iirc

https://warwickts.com/3997/Pomona-2269-Double-Stacking-Banana-Plug

but you want binding posts, for convenience, imo

it would drive me insane to use shielded banana jacks or BNC for anything related to circuit boards on a resistor box

[coppercone2]

https://www.newark.com/pomona/4243-0/double-binding-post-15a-black/dp/89F359

2.5kV

[TimFox]

Standard spacing for dual bananas or binding posts is 0.75 in = 19 mm.
If you want to go crazy, try an SO-239 panel female (“UHF”) and a short ground binding post 0.75 in away, as on early Tektronix CROs.  The SO-239 will accept a 4 mm banana plug.  They are available with Teflon insulation.
 BNCs are easier to mate, but TNC, N, and UHF mate more reliably with threaded collars.

[Overspeed]

Standard spacing for dual bananas or binding posts is 0.75 in = 19 mm.
If you want to go crazy, try an SO-239 panel female (“UHF”) and a short ground binding post 0.75 in away, as on early Tektronix CROs.  The SO-239 will accept a 4 mm banana plug.  They are available with Teflon insulation.
 BNCs are easier to mate, but TNC, N, and UHF mate more reliably with threaded collars.

Hello

Thanks for the idea I link two pictures

239 handle 500 V and 1000 Watt of power , nice specifications for power testing as load because banana 4 mm handle 6 amp easy .

Regards
OS

[coppercone2]

That is the better method if you intend to use HV IMO

But still sometimes you don't have a convenient UHF cable so you use bananas 

[Overspeed]

Standard spacing for dual bananas or binding posts is 0.75 in = 19 mm.
If you want to go crazy, try an SO-239 panel female (“UHF”) and a short ground binding post 0.75 in away, as on early Tektronix CROs.  The SO-239 will accept a 4 mm banana plug.  They are available with Teflon insulation.
 BNCs are easier to mate, but TNC, N, and UHF mate more reliably with threaded collars.

Hello

Be careful with the 19 mm when the use is linked banana regarding tolerances and side force on the contacts , 3/4 inch is 19.05 and the two holes shall be drilled with accuracy , when single banana are used there is no need of holes spacing precision. That why some banana have basket design as Pomona in picture

Banana surface contact even gold plated are not very strong regarding wear , even dedicated aerospace noble metals alloy ( Johnson Matthey ) are not very strong .

Regards
OS

[EC8010]

I don't think it makes sense unless its high value enough or low current.

You forgot low voltage. Almost all my stuff is at least one of the three. The ability to insert a resistance box in the middle of a low noise amplifier without suffering massive hum has been very useful for me. And I have a three decade box that goes to 100M.

There are various adapters for going between spaced 4mm plugs and BNC. I find the most useful one is the one that turns a standard BNC plug into two 4mm pins, allowing a BNC cable to plug into a DMM or to take power from a power supply. I now do all of my connections on coaxial cable into screened boxes because it minimises hum. I haven't had trouble with the 3/4" spacing but (oddly) I have had trouble with the BNC end of some adapters.

Finally, beware the difference between 50R and 75R BNCs. The insulator material depends on type, but 50R pins are a slightly larger diameter than 75R pins (to get the right characteristic impedance). The result is that if you accidentally plug a 50R plug into a 75R socket, you can expect intermittent connection thereafter when you plug 75R plugs in because you've splayed the socket's connector. Most lab gear is 50R but broadcast (analogue) video was 75R. I keep my 75R stuff in separate tins well away from the 50R and with terminations painted either violet or green. Domestic S/PDIF audio digits are 75R.

[Overspeed]

I don't think it makes sense unless its high value enough or low current.

You forgot low voltage. Almost all my stuff is at least one of the three. The ability to insert a resistance box in the middle of a low noise amplifier without suffering massive hum has been very useful for me. And I have a three decade box that goes to 100M.

Finally, beware the difference between 50R and 75R BNCs. The insulator material depends on type, but 50R pins are a slightly larger diameter than 75R pins (to get the right characteristic impedance). The result is that if you accidentally plug a 50R plug into a 75R socket, you can expect intermittent connection thereafter when you plug 75R plugs in because you've splayed the socket's connector. Most lab gear is 50R but broadcast (analogue) video was 75R. I keep my 75R stuff in separate tins well away from the 50R and with terminations painted either violet or green. Domestic S/PDIF audio digits are 75R.

Hello

3 parameters are used in connectors rating
Max voltage : as that linked with insulator performance
Max current : as that linked with the contact surface
Max power : for heat dissipation / cable size

Low voltage can be high current and 239 connector are far better than the small BNC , 239 are perfect for dummy load or shunt .

Regards
OS

[EC8010]

I figure we don't need to worry about a connector's voltage, current, or power rating when it's being connected to a decade resistance box; the resistance box is a measurement instrument (not a dummy load as students think), so we're always looking at little signals. Doubtless, someone will come up with an example that contradicts. 

[coppercone2]

People do it, that's why there is cheap fried boxes around


I am guessing its things like to figure out a bias resistor. I was using the cheap one that you 'dial' to load my HV current source lol.
What better load is there to determine the linearity of a current source??


I noticed, 5 transistors later, that its really easy to switch it to the wrong decade and get a 1000% overload.

[Overspeed]

I figure we don't need to worry about a connector's voltage, current, or power rating when it's being connected to a decade resistance box; the resistance box is a measurement instrument (not a dummy load as students think), so we're always looking at little signals. Doubtless, someone will come up with an example that contradicts. 

Hello

Yes and no , for current source test sometime current reach 1 Amp even if that more common to use a 10 ma source , but for measurement and calibration purpose I agree with the fact to stay in low level current , most 4 wires setup on multimeter use 1 mA current.

Regards
OS

[coppercone2]

and you can also overload it with something simple and stupid like trying to figure out a current limiting resistor.

Ideally you would have a rheostat, but who has a series of those laying around?

[EC8010]

Resistance boxes are a bit like curve tracers; completely capable of destroying themselves and whatever they're connected to. They're so fundamental that it's easy to make a mistake and your thinking hat needs to be on at all times. On the plus side, they often go for peanuts because they've been damaged, so that's handy because it doesn't usually cost much to repair them.

[Overspeed]

Resistance boxes are a bit like curve tracers; completely capable of destroying themselves and whatever they're connected to. They're so fundamental that it's easy to make a mistake and your thinking hat needs to be on at all times. On the plus side, they often go for peanuts because they've been damaged, so that's handy because it doesn't usually cost much to repair them.

Hello

The rule is use the right tool for the right job and don't overkill , no need a metrology class resistance box for daily bench work except if you work on specific instruments
That a mandatory need to own at least two level of resistor box quality / precision , one basic easy to repair in case of mistake one as reference for calibration or dedicated tests , that the same with single resistor box .

I have a DIY resistor box checked with my 6.5 digit DMM base on 0.1 % 50 ppm resistors purchased on Ebay

Connectors are another concern as there is mechanical wear what ever they are gold plated and wiring also if the setup shall be ''perfect'' , there is interesting YT video on Marco Rep channel .

As I am not ''rich'' and as also I don't want to spread / spoil money in unrealistic equipement I try to be clever , BNC is an fordable solution as BNC are available from surplus , 239 connector also

Regards
OS

[tooki]

Yes for PTFE , I have read that 75 Ohms have PTFE and most of the 50 Ohms have nylon ( PA66 ) but that possible to locate BNC 50 Ohms with PTFE , I will check in my available batches

I don’t think one can make such a generalization. 50 ohm BNCs use a wide variety of dielectrics: polypropylene (PP), PBT, nylon (PA), Delrin (POM), polymethylpentene (PMP, TPX) and PTFE were just what I found by looking at a couple of standard, low-cost (under $5) BNC connectors on DigiKey.

And conversely, it’s easy to find 75 ohm BNCs with dielectrics that aren’t PTFE.

The only concern I have seen on BNC is the maximum possible matting before contact surface damage as from main brand as Amphenol they rate at 500 connections before possible contact damage ( galling / fretting)

I would expect contact deformation to be a bigger issue than galling or fretting.

But regardless, I mean… BNCs are used on test equipment here and there. Like, say, every oscilloscope, function generator, and bench LCR. And tons of broadcast video equipment. Surely if BNCs lacked sufficient robustness, they would not be used on such a huge variety of equipment that is plugged and unplugged constantly.

If you use decent quality connectors, connector life should not even remotely be a problem.

[Overspeed]

Hello

As I have not succeed to remove / solve the quote ''mix''  I have cancelled the post and I post a new one

On point 1
Yes and No , if the connector is from a well know brand with a engraved type on , that easy to check what are the specs , when I speak of BNC that instruments quality BNC , they was on aerospace equipment dismantled for recycling . Most of them are Radiall brand
For ''other'' I agree and more if that ''aliexpress '' supply that just gambling
On point 2
I have run a lot of surface analysis from relay to Mil connector and other contacts for aerospace purpose , galling / fretting / wear is a real problem even on top the range brand as Glenair ( check the price on Mouser $$$$$ )
Gold plating is not homogenous when done by electroplating , only PVD coating is and even by using PVD , gold is not a hard metal at all , best is to use Osmium as use on probe tip for IC /wafer testing , you can use Ruthenium which is better than Osmium.
 
But the worth case for contact is friction as in slip rings or test connectors

I link a old but interesting doc on sliding metal for contact

Regards
OS