Sloppy front panel BNC connector - 475


Torch
 

The 475 has the 3-conductor BNC terminals on the front panel, to indicate the correct voltage/div when a compatible 10x probe is connected.

Channel 1 on my 475 became very positional, by which I mean there was no signal unless the connector was "just so". Wiggle it a bit and the signal would drop out.

I opened it up, removed the shield from the attenuation module and found one lead of the little 43 ohm resistor was snapped off at the body. A closer examination revealed that this was because the BNC connector proper is slightly loose inside the outer, 3rd conductor, ring. At first I thought that perhaps the nut came loose, but no, the nut is tight, the play is withing the connector itself. Over time, I guess metal fatigue set in, particularly since the resistor lead was cut down as short as possible.

Now, I can replace the resistor. It's a bit of an odd value, but looks like a 1/8 watt, 5% carbon film component. However this will likely happen again unless something is done about the BNC connector. I assume it is proprietary. Does anyone know of a source or part number for this thing? Alternatively, does anybody know of a fix for this? Crimp the outer conductor slighly? Apply some cyanoacrylate and let it ooze into the joint?


Tom Lee
 

I've noticed the same problem in more than one scope, and considered various options similar to your list. Ultimately, lazybones that I am, I've simply added a short length of stranded wire to the resistor. That allows the BNCs to continue flexing a little without imposing stress on the resistor. If you can't figure out a better alternative, consider this workaround as a Plan B.

-- Cheers,
Tom

--
Prof. Thomas H. Lee
Allen Ctr., Rm. 205
350 Jane Stanford Way
Stanford University
Stanford, CA 94305-4070
http://www-smirc.stanford.edu

On 1/28/2021 22:57, Torch wrote:
The 475 has the 3-conductor BNC terminals on the front panel, to indicate the correct voltage/div when a compatible 10x probe is connected.

Channel 1 on my 475 became very positional, by which I mean there was no signal unless the connector was "just so". Wiggle it a bit and the signal would drop out.

I opened it up, removed the shield from the attenuation module and found one lead of the little 43 ohm resistor was snapped off at the body. A closer examination revealed that this was because the BNC connector proper is slightly loose inside the outer, 3rd conductor, ring. At first I thought that perhaps the nut came loose, but no, the nut is tight, the play is withing the connector itself. Over time, I guess metal fatigue set in, particularly since the resistor lead was cut down as short as possible.

Now, I can replace the resistor. It's a bit of an odd value, but looks like a 1/8 watt, 5% carbon film component. However this will likely happen again unless something is done about the BNC connector. I assume it is proprietary. Does anyone know of a source or part number for this thing? Alternatively, does anybody know of a fix for this? Crimp the outer conductor slighly? Apply some cyanoacrylate and let it ooze into the joint?




Ed Breya
 

The main thing for proper HF response is that the outer BNC shell is properly grounded to the chassis, so the input current returns to ground and the low end of the probe cable. The probe coding ring structure can be loose without affecting the HF performance.

The little resistor in series with the center conductor is actually quite important. It serves as a damper in part of the transient response, and more importantly, it is a fuse against severe input voltage excess.Ideally, if an excess voltage far beyond the normal input rating is applied, and is of unlimited power, the resistor should literally explode and disconnect the input circuitry from the connector. The unlimited source rating is typically up to 1500 Vrms, possibly more, and the remains of the resistor should be able to withstand this after clearing the arc. It does not protect from damage, only from igniting a fire, according to various standards.

In the old days, it was typically a 1/4 to 1/10th W Allen-Bradley carbon composition resistor. The exact size and rating don't matter too much - with enough input fault, the body will quickly disintegrate, and the leads will burn away until the arc is extinguished. In more modern designs, the resistors often were changed to carbon film (on a ceramic core), and even board-mounted, or even surface mount, for lesser rated inputs. I have never trusted any of these alternatives for general purpose scope inputs, so I'd recommend always using good old carbon composition resistors in such spots - they make the best fuses of any type resistor, when suspended in mid air.

In reality, it's not likely that our scopes will be exposed to such extremes, but it's easy to go with CC resistors and be sure it's as good or better than original, in this respect.

Ed


Tom Lee
 

Hi Ed,

I misunderstood the original post. I thought he was talking about the resistor for the scale factor switch, not the series damping resistor on the center conductor. But thanks to your post, I see that I missed the "43 ohms" part of his description; the probe coding resistor is 100 ohms.

Thanks for your eagle eyes.

And I second your recommendation to keep it a carbon comp if replacement is needed.

-- Cheers,
Tom

--
Prof. Thomas H. Lee
Allen Ctr., Rm. 205
350 Jane Stanford Way
Stanford University
Stanford, CA 94305-4070
http://www-smirc.stanford.edu

On 1/29/2021 01:38, Ed Breya via groups.io wrote:
The main thing for proper HF response is that the outer BNC shell is properly grounded to the chassis, so the input current returns to ground and the low end of the probe cable. The probe coding ring structure can be loose without affecting the HF performance.

The little resistor in series with the center conductor is actually quite important. It serves as a damper in part of the transient response, and more importantly, it is a fuse against severe input voltage excess.Ideally, if an excess voltage far beyond the normal input rating is applied, and is of unlimited power, the resistor should literally explode and disconnect the input circuitry from the connector. The unlimited source rating is typically up to 1500 Vrms, possibly more, and the remains of the resistor should be able to withstand this after clearing the arc. It does not protect from damage, only from igniting a fire, according to various standards.

In the old days, it was typically a 1/4 to 1/10th W Allen-Bradley carbon composition resistor. The exact size and rating don't matter too much - with enough input fault, the body will quickly disintegrate, and the leads will burn away until the arc is extinguished. In more modern designs, the resistors often were changed to carbon film (on a ceramic core), and even board-mounted, or even surface mount, for lesser rated inputs. I have never trusted any of these alternatives for general purpose scope inputs, so I'd recommend always using good old carbon composition resistors in such spots - they make the best fuses of any type resistor, when suspended in mid air.

In reality, it's not likely that our scopes will be exposed to such extremes, but it's easy to go with CC resistors and be sure it's as good or better than original, in this respect.

Ed




Torch
 

Interesting about the resistor material. The original definitely looks like carbon film, and the service manual confirms that (well, it states that resistors used in this model with 4 bands are carbon film, 5 bands are metal film).

Audiophiles claim that carbon composite resistors are noisy. Does that also apply to electrical noise in the scope front end?


Roy Thistle
 

On Fri, Jan 29, 2021 at 07:01 AM, Torch wrote:


carbon composite resistors are noisy. Does that also apply to electrical noise
in the scope front end?
It's like 53 ohms... At 1 Hz bandwidth, and room temperature, that's about 1 nV. (possible more about carbcomp, in a later post)


Torch
 

Lacking a 43 ohm resistor in my current stash, I soldered two 22 ohm resistors in series temporarily to verify there is no issue with the BNC ground side. That part is OK, no loss of signal when I wiggle the connector, but there is definitely more ringing looking at a fast pulse. I suspect this is due to the additional inductance caused by stringing two resistors together. Would not the additional length of adding a stranded wire cause a similar artifact?


SCMenasian
 

The excess (over thermal) noise due to a resistor is proportional to current (and, in some cases, possibly even a power of current). If there is no DC current flowing (as is, I think, your case), only AC current would contribute. A large AC current implies a large signal; so, the S/N ratio, even for a carbon composition resistor, would be large and the excess noise of a carbon comp resistor over a metal resistor would probably be insignificant. The current path in metal film resistors is often a spiral cut in the film, at least for high resistance values; inductance could be a problem at high frequencies - I don't know about 50 ohm resistors, though.


saipan59 (Pete)
 

See also my recent post on my 2465:
https://groups.io/g/TekScopes/topic/79720337#176586

As you described, the inner part of the BNC was slightly loose when the center contact was broken, but the nut was tight, and no visible damage.

Pete


Roy Thistle
 

On Fri, Jan 29, 2021 at 08:11 AM, Roy Thistle wrote:


more about carbcomp, in a later post
Of all the various resistor technologies (excepting perhaps specialized ones), carbon composition resistors have the best response, at least, to peak pulsed power (like a one-time high power transient)... supposedly not only due to their thermal bulk; but, also in the way they conduct and dissipate thermal energy. Other types, are not "as good."
Where "as good" means of course, ceteris paribus, and caveat emptor and all that.


Roy Thistle
 

On Fri, Jan 29, 2021 at 09:51 AM, SCMenasian wrote:


noise due to a resistor is proportional to current
Hi... can you say more about that? ... are you talking about 1/f aka flicker aka pink noise... because were usually at 10s of MHz on TekScopes.


Roy Thistle
 

On Fri, Jan 29, 2021 at 09:51 AM, SCMenasian wrote:


inductance could be a problem at high frequencies
One could see the Vishay PR02-FS series, billed as "High Pulse, Fusible, Non-Inductive, Flameproof, Metal Film Resistor"
But I'm not claiming anything about them in the application discussed in this thread.


Torch
 

So the set screw described in your thread as holding the BNC in place, was this holding the BNC part to the outer ring, or securing the entire thing to the chassis?

It sounds promising, if all I need to do is tighten a set screw!


Tom Lee
 

Yes, adding inductance will degrade transient response. As I mentioned in a follow-up post, I thought erroneously that you had been talking about the other resistor -- the one that illuminates the scale factor light, and which therefore is outside of the signal path. For the center post, you would have a bit of a challenge finding a short enough stub that also has sufficient flexibility to tolerate wiggling. I don't recommend my lazybones method for that resistor, although you might be able to get away with some bodge.

As to your later question about materials and noise, carbon comps do exhibit large 1/f noise, but the power spectral density of that excess is a function of dc bias current. There is no dc bias current in that part of the circuit, so the noise of a carbon comp will be pretty much just that of any other resistor; it'll just be plain old thermal noise (below 1nV/root Hz for that 43 ohm resistor).

-- Cheers
Tom

--
Prof. Thomas H. Lee
Allen Ctr., Rm. 205
350 Jane Stanford Way
Stanford University
Stanford, CA 94305-4070
http://www-smirc.stanford.edu

On 1/29/2021 08:26, Torch wrote:
Lacking a 43 ohm resistor in my current stash, I soldered two 22 ohm resistors in series temporarily to verify there is no issue with the BNC ground side. That part is OK, no loss of signal when I wiggle the connector, but there is definitely more ringing looking at a fast pulse. I suspect this is due to the additional inductance caused by stringing two resistors together. Would not the additional length of adding a stranded wire cause a similar artifact?




SCMenasian
 

I wasn't being specific about the type of noise; my point was that noise generated by this resistor in this circuit is probably completely insignificant. The resistor appears to be between the input and a high impedance attenuator. The current through the resistor is so low that no type of excess (above thermal) noise generated by the resistor would be significant. Shot noise (which is current dependent) would amount to much less than a nanovolt in the worst case - much less than the input noise spec of the typical oscilloscope. Other, material dependent, types of noise might be present in a carbon resistor; but they are probably also not significant in this circuit.


Tom Lee
 

Hi SC,

You were very clear in your first comment, and of course absolutely correct. The bias-dependent excess noise of a carbon comp. is a non-issue in this case, ‘cause there’s no bias.

—Cheers,
Tom

Sent from my iThing, so please forgive typos and brevity.

On Jan 29, 2021, at 1:37 PM, SCMenasian <scm@menasians.com> wrote:

I wasn't being specific about the type of noise; my point was that noise generated by this resistor in this circuit is probably completely insignificant. The resistor appears to be between the input and a high impedance attenuator. The current through the resistor is so low that no type of excess (above thermal) noise generated by the resistor would be significant. Shot noise (which is current dependent) would amount to much less than a nanovolt in the worst case - much less than the input noise spec of the typical oscilloscope. Other, material dependent, types of noise might be present in a carbon resistor; but they are probably also not significant in this circuit.





Torch
 

I can't see any sign of a set screw. Unless it is hidden under the mounting nut.

Not going to be fun to get that connector out.


saipan59 (Pete)
 

On Fri, Jan 29, 2021 at 05:54 PM, Torch wrote:


I can't see any sign of a set screw. Unless it is hidden under the mounting
nut.
Hi, not sure who you're replying to, but in case it's me (regarding the similar issue on my 2465):
On the 2465, there is no nut for the BNC. Rather, there is a 'collar' that includes a set-screw, and it is only accessible when the attenuator module is removed. I think (not sure) that if the set screw were removed, then the BNC would either slide out or screw out of the attenuator frame.
In my case, the set screw was tight, and the only thing I could see to give mechanical support to the center conductor was the solder joint.

Pete


saipan59 (Pete)
 

Some pix of the the 2465 issue. In one image, on the far left you can see the set screw in the frame.
https://groups.io/g/TekScopes/album?id=260014

Pete


Torch
 

On Fri, Jan 29, 2021 at 08:30 PM, saipan59 (Pete) wrote:


Hi, not sure who you're replying to, but in case it's me (regarding the
similar issue on my 2465):
On the 2465, there is no nut for the BNC. Rather, there is a 'collar' that
includes a set-screw, and it is only accessible when the attenuator module is
removed. I think (not sure) that if the set screw were removed, then the BNC
would either slide out or screw out of the attenuator frame.
In my case, the set screw was tight, and the only thing I could see to give
mechanical support to the center conductor was the solder joint.

Pete
Pete,

Yes, sorry, I was referring to your post. Thanks for the photos, from that I can see the 475 mounting is different. My BNC connectors are secured with a more traditional panel-mount nut. And not much bloody working room around it! ;(