On 05/24/2012 02:36 PM, Fred Cisin wrote: Um. Is that not what the design that's been floating around (the one with the 555s) does? -Dave -- Dave McGuire, AK4HZ New Kensington, PA

Or on tapes where the holes have been stripped for one reason or another... On 5/24/2012 2:07 PM, Al Kossow wrote:

On 5/24/12 3:15 PM, Chuck Guzis wrote: HP, Digitronics, Feranti built friction feed readers. The HP and Digitronics have problems with the pinch wheel going gummy. Never poked inside a Feranti.

Hi Fine except no holes is data as well.You really have to have a wheel giving the approximatedistance.Dwight

It occurs to me that for recovering tapes with problematic feed holes (which are used for clocking, even if they are not used for feeding), the ideal would be an optical reader, which fed via a capstan, and feed at a good, *relatively* constant speed (or perhaps even via a stepper motor). One could then use the data itself along with the clocking from the sprocket holes to recover the data, even for stripped or tapes with uneven holes - perhaps with some manual assistance. One could save timing information from the data holes AND the sprocket holes for later processing, analysis and recovery. But for any one person, probably more trouble than it is worth. I have a few tapes with bad sections, but those could always just be edited and fixed manually, after reading in the whole tape. Jay On 5/24/2012 2:07 PM, Al Kossow wrote:

I don't know much about paper tape (I never saw it in person), but I'd use just a good old microcontroller (like an Atmega 8, as an example) and do all processing inside, to spit IEEE488, serial or parallel ascii data.

On Thu, May 24, 2012 at 3:01 PM, Dave McGuire <mcguire at neurotica.com> wrote: Most likely. The design from Byte happens to take advantage of asymmetry in the phototransistors (at least in the parts they had at the time) by installing the sprocket sensor in an orientation that gets light on it later, to give the data holes a little extra time over the data phototransistors before the latch fires. Now that I have a liquid-resin 3D printer to play with, and with upper surfaces that are parallel to the exposure plate having a natural smooth finish, I was contemplating making a printed shape that is essentially a 2"-long block with a 1"-wide "groove" and a line of 9 holes to install phototransistors under. Unfortunately, at around $1/g of resin, it's probably cheaper to send a part like that out to a machine shop and have it milled out of Delrin. -ethan

Be warned that using machien tools well takes a lot of practice. You will not be able to amke a tape reader in your first week of using such devices :-) The traditional way is to use a capstan and pinch roller, as in a tape recorder. This gives constant speed, of course. The captan normally turns continuously, the pinch rolelr is engaged by a solenoid to start the tape moving. There was normally a brake, nothing more than an electromagnet on one side of the tape and an amrature plate on the other side that grips the tape when the electromagnet is energised, to stop the tape. As I siad, unless you require start-stop operation, it can be simplified a lot. Possibly an 2748? I'ev got one on my HP9800s. There are a few pictures of it on the HPCC ewb site... Yo ucan get the HP service manaul from the Australian museum site. That is actually not a bad manual to read if you wwant to make a start-stop reader, in that it descibes the ,echanism. One odd thign about the HP one IIRC is that the tape motion is controleld by a magnetic clutch betwee nthe mtoro and capstan, and not by engaging the pinch roller. Note that if you dismantle this part of the HP2748, you need a special alignmnt tool to put it back togetehr or the clutch will wear out vey quickly. Making that sool is not a bad project if you have a small lathe (should only take half an hour or so). I didnt think it was _that_ fast. -tony

IIRC, the HP is somewhat unconventional here. The pinch roller is controlled by a solenoid, but it is only off-tape dureing loading. When the machine is in operatin, the pinch rolelr is enganged. The tape motion is controlled by an electromagnetic clutch betwee nthe (shaded pole) motor and the capstan.There is the normal brake design. I managed to find a spare read head for mine on Ebay (totally mis-descrived, which it probably why nobody else bid). The reader heads can be taken apart and repaired (I've done it), but it;s ntot all that easy. The main problem is that the plate carrying the phototransistors has to be electrically isolate dfro mthe chassis, and there are various 'paper' insualtors you haev to get lined up. The trick is to make 2 metal rods 1/8" diameter and about an inch long. Put those into 2 blind holes i in the lower body of the read head (these hols are there solely for that purpose), then fit the insualtors and phototransistor plate, fit the screws and tightn up, then remove the pins. Fit the PCB and solder all the wires in place. As I said I have one, and I've taken it all apart and repaired various bits. The main problem with mine was that the bearing in the magnetic clutch had failed. The official repair is to change the complete clutch. I made up some mandrels and dies and prssed the clutch apart in a bench vice, and discovered the bearing was a stnadard size ball race. I had no problems getting a new one and pressing it back in. Electronically they are very simple. The 3 plug-in PCBs are basically PSU, tape motion control and read amplifiers. I do find the amount of hardwiring under the chassis to be moderately amusing.... Checking, it's faster you you than me. I think you get 500cps out of it, I get about 440 (due to the difference in mains frequency, the motor runs slower over here). That's not bad, although I have things that are faster... I think there's such a reader on Ebay at the momnet, if anyone's looking for one. -tony