My New R to R Project - Ampex VPR-2B

[glimmie]

My new toy. This is a 1982 vintage broadcast video recorder. I know these machines inside and out. I have many concerts and the like copied from edited masters. Some are Dolby A encoded and I have a CAT 363 decoder to go along with it.

This is a four motor transport using brush type DC motors, two reel motors, the capstan, and video head durm. Each motor is under servo control. The tape tension is controlled extremely accurately - it has to be to meet broadcast specs. These servo systems are also interlocked to each other. The capstan is belt driven as direct drive is not as smooth. But Ampex did abandon the belt drive on later models.

Audio performance is quite good, three channels 50-15K but it can go a bit wider. 50-15K was/is NTSC TV proof of performance. A lot of these concerts are mixed down from 16 or 24 track direct to the video master so they are on par with studio audio masters.

However there is a drawback to really good audio. The complex servo and control systems generate all kinds of clocks and other pulses right inside the audio spectrum. The power supplies are also exposed to this junk. So I plan to do an external tube based preamp just like the Tape Project. The internal audio system is good, but an outboard tube preamp should be a welcome improvement.

These machines are "direct color high band" recording. This is leaps and bounds above the home VHS format. The machine captures the entire NTSC spectrum from DC to 4.2mhz at around 48db video S/N. Unlike DVD it is also an uncompressed analog recording. Uncompressed video has a distinct quality difference from DVD just as LPs and analog tape has from CDs. And while most movies are transferred to DVD from film via component video, these live concerts are forever locked into the NTSC composite format. So my tapes will always be superior to a DVD of the same concert as the DVDs are made from the same source but are also compressed video.

For those not familiar with analog video recording, the raw video signal is not simply mixed with a bias and put on the tape as is the audio. That would require a 17 octave spread on tape with the all familiar 6db per octave loss. That means an equalizer with a 106db gain spread - impossible to make! So the video modulates an FM carrier from 7.09mhz to 10mhz. This is now less than an octave spread and can be easily equalized. Of course we still need a head to tape speed of 1000ips to lay down a 10mhz signal on the tape. VHS and Betamax machines used a much lower FM frequency due to the much slower head to tape speed - about 400ips. So in turn the incomming video must be rolled off to less than 3mhz thus substantially reducing image detail. Color must be seperated in VHS as well and recorded at an even lower frequency which creates the poor S/N ratio we had with VHS.

[ironbut]

That is too cool glimmie (btw we use our real names here, you can just add it in your signature in the control panel)! It sounds like you know your video stuff pretty well. Me? I don't know my ,.. well, you get the idea. I'd love for you to give us a little run-down of the pro video format(s). Stuff like tape speeds, widths and what the heck "helical scan heads" is all about. Actually, that might be a little involved but I don't think that most of us even know what is being used in our local news network.
I'm sure you've seen this but here's a little history page on Ampex video;
http://videopreservation.stanford.edu/museum/
I was looking for that site that has the story of transferring those moon landing videos. That was a really interesting story. Do you have that at hand?

[glimmie]

Well here is an excellent page on VTR theory and history for anyone interested.

http://www.lionlamb.us/quad/theory.html

As for the "helical scan heads" the video heads are on a spinning drum that writes a slanted (helical) track across the tape. The big 2inch monsters from the 60s and 70s were known as "Quad" as they wrote almost vertical tracks on the tape by alternating four heads. Helical machines were first used in non broadcast applications due to stability issues with the long videotracks. By the late 1970s technology caught up to where a broadcast helical machine could be realized. Even today the high end HDTV broadcast cassette VTRs still use a helical scan system not that far off from consumer VHS mechanics. Just as with audio, open reel video tape is just about extinct.

As audiophiles well know, the faster the tape speed, the more HF we can record. So 15ips have much cleaner and powerful high end than does 7.5ips. Well in video we need to record in the mhz range. Well even 30ips comes up way short of this task. So we need to get the tape speed up. We could run the tape at 300ips or faster and some early VTR attempts tried just that. However Ampex invented the idea of spinning the video heads against a traditional 7.5 or 15ips linear tape speed to "cheat" the realitive tape speed up. Of course you now have a whole new set of problems. The track is no longer a linear endless (for the program duration) track. It has a start and end point. The tape must be positioned so that the video head contacts the track at the proper time. This is where the servo system comes into play.

Another tidbit: Ray Dolby was part of the origional design team at Ampex that built the first VTR in 1956. He was directly responsible for the multivibrator type of FM modulator which greatly simplified and improved the video record chain. And most people thought he was an audio guy!

[Ben]

Some quick comments
I remember reading in some old magazine, about a DIY video recorder
project around the mid 70's. They did recording the hard way ... very fast tape speeds. I think you got like 20 seconds off a 4800 ft tape.
I remember watching 1" B&W video tape in the early 1980's. That must be a very extinct format!

[ironbut]

Great explanation Andy. I can see now why some video reels were 14". I did have the pleasure of seeing the full restored VRX1000 up at Stanford last year. It's the first machine from the relocated Ampex Museum to be ready for display. It's about the size of two frig's with one standing up and the other laying down.
Do you know if they ever experimented with using air pressure instead of a capstan/roller like they did on some main frame computers?
Really fascinating stuff, thanks for sharing.

[glimmie]

"Do you know if they ever experimented with using air pressure instead of a capstan/roller like they did on some main frame computers?
Really fascinating stuff, thanks for sharing."

Absolutly. Ampex was the only company with an all air bearing transport and a vacuum capstan. The VPR-3 which succeded my machine used air guide technology. It had incredible cueing and tape handling performance. Apparently you could grab a reel in fast wind and not damage the tape! I don't think any end user ever dared to try it though.  See one in operation here:

http://www.youtube.com/watch?v=-z7ZAZDgiMg

Before that Ampex built the AVR-1 which used vacuum colums to regulate tape tension. Another defunct company, IVC made the IVC9000 VTR which also was a vacumm colum design.

The old 2in machines very early on went to an air bearing on the video head scanner. In the 2in Quad format the speed of the heads is 14,400rpm. In the early 1960s they did not have digital memory to iron out the jitter as we do today so they had to get the mechanical stability as high as possible. Ball or sleave bearings did not cut it.