Back in the Game!

Forum for discussion of narrow-bandwidth mechanical television

Moderators: Dave Moll, Andrew Davie, Steve Anderson

Postby gary » Sat May 25, 2013 6:29 pm

Yes Ralph, without the bias resistor the charge across the capacitor will eventually float to rail. For a dual supply op-amp it is sufficient to just have a bleed resistor to ground but for a single supply op-amp the input needs to be DC biased - the capacitor between ground and the 10k resistor is so the gain will be unity at DC.

I think it is possible to have the DC biasing on the negative side but it would complicate the gain calculation significantly. AND you would still need the bleed resistor. I think if you were to do that you would probably use a proper voltage reference rather than just a voltage divider.

NOTE: that now you HAVE to carefully calculate the RC values so as not to limit the required bandwidth (as it is obviously now an RC filter).

Ralph, I have seen the letter 'em' used to denote micro but it is always upper case M to distinguish it from m=milli.

BTW I think this is a perfect example of where you would be wise to (if necessary) reorganise the preceding circuitry to allow DC coupling.

In addition the bias circuitry makes this arrangement have a worse performance than a standard ac coupled inverter (noise, lower input impedance, etc) so if the other half of the LM386 is spare it might be better to use that as an inverter (I am thinking out loud there).
gary
 

Postby Steve Anderson » Sat May 25, 2013 9:11 pm

OK, seems a bit of confusion has crept in here. I assumed that the source for this would be AC-coupled, if it's audio gear it is 99.9% certain it will be.

The LM358/324 etc are designed as single-supply op-amps, the input can go all the way to ground, feedback loop allowing, in fact to around 200mV negative. So with AC-coupling into the 'contrast' pot parts of the signal will be well below -200mV, so it needs 'lifting up' via said voltage divider to keep the input signal within the linear range of the op-amp.

Those bias resistors also provide a path to ground for the input bias current (Typ. 20nA) otherwise with the 1uF cap they have nowhere to go and the output will simply sit at one rail or the other depending on configuration., The input to these op-amps is a PNP transistor, so current actually comes out of the input terminals...even though it's called an Input Bias Current, it's given a -ve value, in this case typically -20nA. There is no need for the 10k resistor on pin 5.

An op-amp tries to perform the gymnastics required to keep the +ve and -ve inputs as equal as possible, if, as here, the +ve input goes slightly positive the dog will wag its tail to try and get the -ve input the same...this is how you get gain from the ratios of the two feedback resistors.

Now without the suggested 10uF cap the op-amp has a gain of around 6, even at DC, it will attempt to amplify the +6V bias by 6, but 36V isn't available! It will sit at the positive rail. The 10uF reduces the gain at DC to unity, 6V in, 6V out. The value is chosen for a -3db point of around 1Hz.

Many trip up on bias currents, even when they would appear insignificant, some op-amps are measured in fA (femto-amp), one thousandth of a pico-amp, but even so, that current needs to go somewhere.

The input impedance of this stage is 10k (the pot) two 220k in parallel across 10K drops it fractionally, the op-amp is megs...

Steve A.

You don't have to use a single-supply op-amp in this configuration, even a grotty old 741 will do the job, no circuit changes required...but I don't suggest using a 741...please, something better.
User avatar
Steve Anderson
"Fester! Don't do that to 'Thing'"
 
Posts: 5361
Joined: Fri Mar 30, 2007 10:54 pm
Location: Bangkok, Thailand

Postby Ralph » Sun May 26, 2013 12:56 am

Steve and Gary - thank for the tutorial. An ample supply of LM358's will arrive early in the week, so some bench time should result in a solution. Will post the results.....

Ralph
User avatar
Ralph
Evil Genius
 
Posts: 101
Joined: Fri Sep 07, 2007 2:39 am
Location: Mason, Michigan USA

Postby Steve Anderson » Sun May 26, 2013 1:18 pm

Hi Ralph, it's said that a picture is worth a thousand words, to me a diagram even more. So here's a quick fag-packet sketch of what I suggest which hopefully clarifies the text in previous posts.

If you should have problems with supply noise getting into the signal (tut-tut) the revised bias arrangement as shown in the inset should fix that, good luck.

Steve A.

I don't usually show bypassing/decoupling capacitors in my diagrams, they're implied as part of good practice.
Attachments
LED Drive 1.gif
LED Drive 1.gif (8.9 KiB) Viewed 21667 times
User avatar
Steve Anderson
"Fester! Don't do that to 'Thing'"
 
Posts: 5361
Joined: Fri Mar 30, 2007 10:54 pm
Location: Bangkok, Thailand

Postby Ralph » Mon May 27, 2013 1:51 am

Steve,

My problem was that I had almost always done AC-coupled applications via the inverting input, biasing the non-inverting for any needed offset. It never occurred to me that I still bias the non-inverting input for a fixed DC offset (DC gain of unity) while merrily amplifying an AC signal at the same input,

Now that I have your schematic (and a nice on-line tutorial on op-amp biasing from MIT), all should be well. The TI application notes were woefully lacking in in terms of differentiating AC vs DC service - especially with respect to the non-inverting input! Most of the mods can be accomplished with the new network tacked on the back side of the PC board, so it shouldn't look like major surgery.

Thanks all-around......

Ralph
User avatar
Ralph
Evil Genius
 
Posts: 101
Joined: Fri Sep 07, 2007 2:39 am
Location: Mason, Michigan USA

Postby Steve Anderson » Mon May 27, 2013 12:58 pm

Ralph wrote:The TI application notes were woefully lacking...Ralph

Semiconductor datasheets and application notes vary widely across the industry. The best I find are from ON Semiconductor, Philips, National Semiconductor, and Microchip, most of the rest are 'OK', sort-of.

The worst are those from the far-east, notably China and Taiwan. Two pages on the environmental impact, four pages on how it's packed, two pages of physical dimensions. Electrical specifications...limited...with statements like, "Input Bias Current...2.2uF" Very useful. And nothing about how to use the device.

Steve A.
User avatar
Steve Anderson
"Fester! Don't do that to 'Thing'"
 
Posts: 5361
Joined: Fri Mar 30, 2007 10:54 pm
Location: Bangkok, Thailand

Postby Ralph » Wed May 29, 2013 2:57 pm

Thanks for all the help guys. Properly biased, the input op-amp functions perfectly and lets me display full dynamic range from virtually any NBTV source. All the televisor circuits are now on a single board that is housed in a compact enclosure (B below) that connects to the televisor mechanics with a simple 4-conductor cable. The mechanics assembly (A below) has very little wiring, with inputs for 12 and 24V power supplies and the 4-conductor interface cable. A small set of sample images (C below) shows (L to R) Bugs Bunny from my Romscanner, A test attern from the Club CD, and the test pattern from Darryl Hock's standards converter.

Now I can get to the polishing for the mirror screw knowing that the same controller electronics will be compatible with both the Nipkow and mirror screw televisors.

Ralph
Attachments
televisor04.jpg
televisor04.jpg (231.16 KiB) Viewed 21602 times
User avatar
Ralph
Evil Genius
 
Posts: 101
Joined: Fri Sep 07, 2007 2:39 am
Location: Mason, Michigan USA

Postby gary » Wed May 29, 2013 3:06 pm

Very nice crisp images Ralph - well done!

I realise it's a difficult thing to get right but a video of the disk in operation would be, I am sure, of great interest to us all on the forum.

Cheers.
gary
 

Postby Ralph » Wed May 29, 2013 11:06 pm

Gary,

My wife is now expecting a peaceful week in the country and those expectations include a complete absence of spinning discs and flashing lights. I will see what I can accomplish upon my return!

Ralph
User avatar
Ralph
Evil Genius
 
Posts: 101
Joined: Fri Sep 07, 2007 2:39 am
Location: Mason, Michigan USA

Postby gary » Wed May 29, 2013 11:15 pm

Have a good'un Ralph - you (and no doubt the missus) have earned it.
gary
 

Postby Ralph » Fri May 31, 2013 4:27 am

Gary,

Quick first-pass having never done this!

Ralph
Attachments
ralphs_televisor.avi
(5.76 MiB) Downloaded 828 times
User avatar
Ralph
Evil Genius
 
Posts: 101
Joined: Fri Sep 07, 2007 2:39 am
Location: Mason, Michigan USA

Postby gary » Fri May 31, 2013 10:43 am

Excellent Ralph! Very nice pictures indeed! A great "first-pass".

Two observations:

These still pictures are very clear and sharp. *Moving* pictures give a perception of even greater clarity - so we should be in for a treat when you get around to doing one of those.

I see you are still getting a bit of hunting. You should be able to "tweak" that out - there is no reason you shouldn't be able to obtain a rock steady picture eventually. So again - something to look forward to...


Again - very well done - and thanks for posting the video.

Cheers!
gary
 

Postby Ralph » Fri May 31, 2013 11:57 am

Gary,

Yes, sometimes it seeks and sometimes it doesn't. Still dealing with three interacting factors (signal sync pulse, distance of the photo-sensor from the strobe pattern, and the biasing of the speed pot) and I haven't even played with tweaking the electrolytic! Also, everything gets amplified by the extreme close-up, but I would prefer the larger size for quality estimation. Also, once I get it dialed in so it doesn't appear to be seeking in the video I know it is rock-steady visually.

I had to get a bunch of hardware together to do the video and then there was some file manipulation as the original AVI file ws 58 mb! Next time around I will use some full-motion material (a Betty Boop cartoon, a Bing Crosby video, and some material from my standards converter), but simply ran out of time.

Still early days, but the basic process turned out to be easier than I thought it might. Thanks for the encouragement....

Ralph
User avatar
Ralph
Evil Genius
 
Posts: 101
Joined: Fri Sep 07, 2007 2:39 am
Location: Mason, Michigan USA

Re: Back in the Game!

Postby Ralph » Thu Nov 13, 2014 2:31 am

All,

Over a year ago my wife began her third battle with cancer and I put all my NBTV projects on hold. Things have not gotten better (they are actually worse), but I got to the point where I really needed some projects to distract me from the daily routine.

Step one was to take the Nipkow Televisor off the shelf and fire it up. It worked well for a few days until the 47 mF capacitor failed. Got that repaired last night and things are working nicely. When last we were discussing the topic, the disc was "hunting" slightly and Gary wanted to see actual motion. If all goes well, I will attach a short video showing my old friend Bugs Bunny without the slight hunting seen in previous videos. The video camera exposure is far from optimum, but you will get the idea.

I am working a a short "talking head" video that should work for Gary's request, but am still working at getting the sound on the video.

I can also report that it helps a lot to have something else to think about for an hour of so each day!

Ralph
Attachments
Nipkow_Bugs_01.wmv
(3.59 MiB) Downloaded 563 times
User avatar
Ralph
Evil Genius
 
Posts: 101
Joined: Fri Sep 07, 2007 2:39 am
Location: Mason, Michigan USA

Re: Back in the Game!

Postby Ralph » Thu Nov 13, 2014 2:36 am

Well it would appear that the video did not upload properly - something else to work on! Looks like I got it this time around!

Ralph
Attachments
Nipkow_Bugs_01.wmv
(3.59 MiB) Downloaded 581 times
Last edited by Ralph on Thu Nov 13, 2014 5:36 am, edited 1 time in total.
User avatar
Ralph
Evil Genius
 
Posts: 101
Joined: Fri Sep 07, 2007 2:39 am
Location: Mason, Michigan USA

PreviousNext

Return to Mechanical NBTV

Who is online

Users browsing this forum: No registered users and 28 guests

cron