CNC video display problem: DFP video cable extension

Oh, a male-male DFP cable has "squeeze locks" on the sides, as opposed to thumb-screw locks. The pins themselves are not the literal pins you see on a VGA cable, but rather on the sides of a central block, less risk of bending. A good pic is here:

altho that top tab appears to be some artifact, or just a pc of tape.

Shorter cables are more available, the problem was long cables, or adapters/couplers. If I could find a fem-fem adapter, I suppose I could just order another cable from haas. Or a male-female cable, to extend the original, rather than replace it.

Show us the mill!

On a sunny day (Thu, 18 Apr 2013 10:01:13 -0400) it happened "Existential Angst" wrote in :

Is the cable correctly terminated at the end? Is it driven by the correct impedance?

If the 26 pin cable is just to feed an LCD directly, that will likely not work due to reflections and cable capacitance. The LCD driver needs to be AT the LCD.

Don't know.

Don't know.

There is a mother-type board in the main electric panel, and then there a board in the pendant/controller (pictured in the mackaay link). The dfp connects the two boards.

That's all I know. So the new cable (about twice as long and about 4x the cross-sectional area, for some reason) works, it just doesn't work well.

If impedance is the primary issue, then conceivably even a longer cable by the same company might prove to be a problem.... altho mebbe less of a problem. Christ, even WIRES are turning out to be complicated.....

With renishaw probing, programmable coolant, rigid tap, macros, blast air,

21 tool carousel, some other stuff. Custom coolant tank from ronco-plastics.com, great company.

The pendant/controller you see there on that arm has been moved to the foot of the machine (toward you), because the machine is next to a wall on that side. A bit of a miracle installation. I am already a legend at Haas.... lol

But this cable/screen problem is a bit of a bummer.

Well, I don't know if this has any relation, but I got some VGA cables years ago, and had horrible blurring of the video. I cut one of the cables open and it was just a bundle of wires in an overall shield! I got disgusted and made my own cables with

5 pieces of RG-178 (I think that's the sub-mini 75 Ohm coax designation). These work great, but are rather time-consuming to make by hand.

I'm guessing these DFP cables are supposed to be twisted pairs for LVDS signals. Either the monitor doesn't have the right termination or the cable is poorly made, and maybe has the wrong impedance, causing reflections.

Jon

So the wires themselves were the wrong impedance?

Some of them say "13 pair".

What does that mean, "right termination"? Someone else mentioned the CABLE not being terminated properly?

Should I measure the ohms? What values am I looking for??

The weird thing is this is a digital cable and the artifacts you describe sound like analog problems. I would suspect that there might be too much voltage drop on the 5V in the cable so the display itself is not getting quite enough voltage to work properly. If you can take the back off your pendant without voiding the warranty, you could measure this.

The Wikipedia article on DFP indicates that it uses four twisted pairs for video data and a few auxiliary lines for DDC. It is a digital signal which makes me wonder why you would be getting analog style ghosting and color bleed though. Are you positive it is actually a DFP interface and not something like analog VGA on a proprietary connector? Can you put an oscilloscope on it and see if the signaling appears to match what's indicated on Wiki?

Anything is possible with CNC controls.

Can you put an

Not easily. Everyone I know "used to have" an oscilloscope.... :( Would low voltage cause ghosting, as anorton suggested? Others have told me (here as well) they've had very poor quality video from crappy cables, exp. extensions, suggesting impedance is a real issue.

Are you able to investigate the OEM cable? Is it molded connectors or screw shell ones that can be opened for investigation? Can you look at the connections at either end where they connect to the control or display to see what's there?

Everyone has had crappy video with crappy cables back when the signals were analog i.e. VGA and older. Digital signaling like DFP, DVI, HDMI, etc. shouldn't show that type of effect with bad cables, it should be more like dropout or wrong pixels due to corrupt data. Since it's differential signaling, it should be pretty immune to cross talk and cable length. If the cable is not using twisted pairs then cross talk could be an issue, but I still think it should manifest differently.

Well, no controlled impedance at all! For best video results, they should be individual coax cables (red, grn, blu plus vertical and horizontal sync) that both provide a controlled (in this case

75 Ohms) impedance plus shielding the signals from each other. Just bundling 6 feet of fine wires in an overall shield is TOTAL CRAP, and wouldn't work well on even the oldest EGA systems from the 1980's. The idea of controlled-impedance is that the signals can be completely absorbed at the receiving end, so nothing bounces back from that end and then ping-pong back and forth a couple times, causing echoes.

Typically, these are LVDS signals on twisted wire pairs, and twisted pair wires generally have a characteristic differential impedance of about 110 - 120 Ohms. The receiving end should have a resistor or active termination circuit to fully absorb the incoming waves to prevent them from reflecting off the receiver and bouncing back and forth on the cable.

You can't measure this easily. You need a fast pulse generator and an oscilloscope, and in this case it needs to all be differential. The method is called Time domain reflectometry. An Ohmmeter will not give a reading of the characteristic impedance of a transmission line, which is what you cable is SUPPOSED to be.

A Norton mentions power, I didn't realize that the LCD was POWERED through the DFP cable as well! That could definitely be the problem. Is there an auxilliary power input for the panel? You could try that with a wall-wart, and if that fixes it, you know it is a power problem. I'm not at all surprised that a 30' cable might drop too much voltage to cause such a problem, and in that case, it could go from perfectly OK to serious garbage with just a slight voltage drop. If there's no other way to power the display, then you might need a better cable with thicker power wires, or as a last choice, to hack your own heavy power wires to bypass the thin wires in the cable.

Jon

I think the big question here is whether this is really a DFP digital video signal, or an analog video signal on a proprietary connector that happens to be the same as DFP.

Impedance might be an issue depending on the construction of the original cable vs. the new one. The thing is that with differential digital signaling impedance shouldn't matter much over a short length like 35' and at the relatively low data rates involved here. More so, an issue causing data errors in a digital video signal shouldn't be manifesting as ghosting and color bleed which sound very analog to me.

Termination should not be an issue since that should be at the end devices, not in the cable and should not have changed since this is just a longer cable connecting the original two devices.

Close pictures of the stuff involved would help answer some of the questions.

35 feet for a video cable? Thats too long. Video cables are short for a reason.

This post is legible. I dunno WTF you did with EA, but..

It seems Apples use these cables (I never heard of 'em), so you might try Apple user groups or forums to see what they use that works well.

The 35ft distance isn't trivial.. you may need a wired transmitter/receiver pair of gizmos, in addition to a proper/better cable.

The other replies are likely correct that the particular cabling used to fabricate this cable assembly isn't suited to your purpose.

Many of Arcade's products are typical China/generic quality, and I guess they thought a long, fat cable was worth $120+.

I don't know where you'd get the cable you need, but it might be very similar to a VGA monitor cable but with those digital connectors.. OR, your monitor really is digital, and the cable you have is a half-assed attempt at a non-compatible product (sound familiar?).

If the folks at Haas haven't been able to tell you what specific type of monitor you have, you probably need to try entering random phone extension numbers when calling. And ask for a pin-out of the connectors.

Just accommodating reading levels, is all.... lol

The price was a knee-buckler, but they went down some. And in this situation, would have been worth it, had it worked. As it stands now, it works but irksomely so.

Actually a dicey situation, cuz then it gets back to the local people, a whole hierarchy/turf who's-doin-what who-knows-what issue.... really a pita....

The most practical thing right now, given the scarcity of this cable type, is to finagle one from Haas, and kluge together a female-female coupler via two female circuit board receptacles. I assume those are available, since the boards are made with them, but that might be another needle in a haystack.

The other poss. is dfp-dvi conversion, and hope the pin-outs stay consistent.

Very enlightening if not sobering thread, appreciate all the input. I have a new respect for g-d WIRES!!!!

Buy a cheap cell phone and call Europe or Asia reps, maybe.

Companies go to great effort to keep customers (who pay the bills) in the dark.

The comment I received from a company "Level II" support rep last week was "those (SD cards) should be (work) fine. The device (was $1200) has a capacity of 6 cards.. there are 4 classes of SD which are like Generation 1 thru 4, all slightly different.

I ordered two adapters, DVI (fem) to DFP(26 pin male), which will plug into each circuit board. Then between them I'll run a DVI cable (m to m). We'll see how that works out. Cables arrive in a cupla days. The adapters were hellishly expensive, $42 each, and the 25 ft cable a whole $10. Just *moving* this effing pendant/controller has so far cost HUNDREDS of dollars, $200+ in cables alone..... goodgawd.....

it sounds like an impedance matching problem, like you'd get if you tried to run VGA or component video through a ribbon cable, cat5, (or other inapropriate cable)

extending the original cable with an identical cable it likely to make the picture no more than twice as bad as the original picture was.

the video degradation you see suggests to me that analogue signale are being sent over this cable, and the original is not a DFP cable, it just has DFP connectors on each end.