Image Map Image Map
Page 4 of 4 FirstFirst 1234
Results 31 to 40 of 40

Thread: Power Supply grounding

  1. #31
    Join Date
    Jan 2007
    Location
    Pacific Northwest, USA
    Posts
    29,553
    Blog Entries
    20

    Default

    I wonder if the transistor base circuit is picking up extraneous noise. You might try doing a little bypass by connecting, say, a 0.01 uF cap between base and emitter on each transistor. That might improve the noise immunity.

  2. #32
    Join Date
    Jun 2012
    Location
    UK - Worcester
    Posts
    2,313

    Default

    >>> I meant to say the 1.0 uF capacitor was placed between the negative +8 VDC of the CP/M machine

    Isn't the 'negative +8 VDC of you CP/M machine' the 0 Volts line?

    Not sure why connecting that to the frame of the CP/M machine (which in turn I assume is connected to protective earth?) should have any effect (other than by accident).

    Your CP/M machine appears to be totally disconnected from the electrical circuits of the flexowriter so (other than 120 Hz on the digital signals from the flexowriter to the CP/M machine and the possibility of electromagnetic coupling beteen the two systems) I don't understand why you are having problems.

    When we bought some new paper tape readers and punches many years ago at work, I had to develop an electronic interface that replicated an electro mechanical whirlybird cam mechanism to interface them to the computer! Much more interesting than some of the cr*p I have to deal with these days!

    Your problems also remind me of the time a colleague put a load of capacitors on a UNIBUS backplane to get rid of a load of mysterious noise that was there. To cut a long story short - the computer was at the same level as the ferry radar in the port next door! The interference would come and go - much like the ferries that caused it .

    Dave

  3. #33

    Default

    Quote Originally Posted by Mike_Z View Post
    Suppose I should have been clearer. I placed the filter cap on the Firden Power supply, right after the rect bridge. It did raise the overall supply voltage. I suppose I should check the change in current in the actuating magnets. Didn't think about that. Thanks Mike.
    The meter you have is likely to not tell you the true RMS of the voltage. Without the capacitor it is likely to read higher or lower than it really is. With the filter cap it will be more accurate. Most meters read average and not RMS but display it as though it were RMS. Only expensive meters have true RMS. I had an expensive 8 digit Fluke and it would only read true RMS with an expensive add-on module. Don't be fooled if the dial says RMS these assume a clean undistorted sin wave.
    Dwight

  4. #34
    Join Date
    Dec 2013
    Location
    Near Milwaukee Wisconsin
    Posts
    1,103

    Default

    That is interesting. I have noticed this also. I have a $20 digital multi meter which acts just like you say. My Simpson 260 seems to work better and my Fluke 83 seems to work the best in these situations. What I'm wondering is how good is an oscilloscope? I use a Tektronix 2230. This scope led me to using the bypass capacitor on the CP/M power supply. During Friden to CP/M testing, I noticed that most any key input only resulted in 50% accuracy. I then ran a test of every key multiple times. I thought that this would point to a bad bit or bits. This didn't. I then used the scope to watch code and clutch magnet signals. This led to improving the sampling timing, but didn't correct the accuracy of reading. This is where I noticed the 60 cycle noise. It was on the Friden's VDC, both polarities and when the Friden was connected to the CP/M computer this same noise appeared on the CP/M 8 volt DC power supply. This voltage was a rough 60 cycle sin wave at about 8V peak to peak. My procedure was not to look for the cause but to attempt to suppress it. Trying a bypass capacitor on the Friden didn't do much. I tried bypass cap's on the opto's inputs, also resulting in little improvement. Nosing around on my I/O board in the CP/M machine I noticed the noise on the 8V power bus. Here I found that the 1 uF cap suppressed the noise to less than 100 mV pp. I have always had trouble with noise on logic circuits and don't claim to know much about it, other than try to isolate the supplies and try to keep the distances short, but that doesn't always work. So suppression becomes my most used tool. I can't figure out why the bypass cap worked much better on the CP/M machine rather than the Friden. It would seem that the capacitor should work better at the source of the noise. Does this suggest that the noise comes from the CP/M machine? Yet connecting this machine to either my IBM-XT or Televideo causes no trouble. Or, connecting it to my Epson FX80 or Teletype Model 43 also causes no trouble. I think that chasing noise like this could drive a guy nuts, especial for a guy like me that is half way there to begin with. So, I'm not going to argue with success, although I realize that this problem may come back on the next project. Maybe I'll be smarter then. Thanks for the help Mike.

  5. #35
    Join Date
    Jun 2012
    Location
    UK - Worcester
    Posts
    2,313

    Default

    In my experience it can sometimes be the act of connecting an oscilloscope up to something that causes the problems in the first place!

    The oscilloscope itself is earthed and can be the source of the mains noise itself.

    It is interesting you say 60 Hz - The flexowriter DC circuits are full-wave rectified 60 Hz - making it (effectively) 120 Hz when you look at it on a scope!

    I remember (back in the bad old days) people deliberately removing the protective earths on their oscilloscopes to avoid this problem. Of course, you now risk killing yourself if a fault develops within the oscilloscope!

    Dave

  6. #36

    Default

    That is why they put preamps at the probe tips. It helps to isolate the ground noise.
    Dwight

  7. #37
    Join Date
    Jun 2012
    Location
    UK - Worcester
    Posts
    2,313

    Default

    Depending on the age and what you have...

    Dave

  8. #38
    Join Date
    Dec 2013
    Location
    Near Milwaukee Wisconsin
    Posts
    1,103

    Default

    Well..... I pretty much have the Friden interface with the CP/M machine. What puzzles me is that the Friden keyboard is permanently attached to the typing machine. So when a key is pressed that key is typed. Whereas with the ASR33, when a key is pressed, the computer echos back that key to the typing machine. Kind of a verification of what key of pressed compared to what key the computer read. This is not possible with the Friden. I have been investigating the use of the Friden with early computers. I found a picture of the British code computer, Colossus. It shows Tony Sales, a rebuilt Colossus and a Friden writing machine. The Friden has the punch and reader removed and probably was only used to make the code paper tape and as an output machine. Friden.jpg
    I also seen some pictures and text stating that the PDP-1 used a Friden. I wonder how it was used? I also found that my Friden SFD-V is a Model 1 Systems Programatic Double-case machine. Mike

  9. #39
    Join Date
    Sep 2006
    Location
    Silicon Valley
    Posts
    1,694

    Default

    Quote Originally Posted by Mike_Z View Post
    I also seen some pictures and text stating that the PDP-1 used a Friden.
    The PDP-1 uses a modifed IBM model B typewriter with a Sorobon keyboard encoder.

    Whirlwind and TX-0 used Flexowriters.
    One thing that some systems did was to change the ribbon color between what someone typed in and what came from the computer.

    Half-duplex with characters locally echoed on the terminal was a common mode of operation well into the 70's
    Univac Exec-8 worked that way, with helpful messages like 'last input ignored' when the system wasn't expecting you to type anything.

  10. #40
    Join Date
    Dec 2013
    Location
    Near Milwaukee Wisconsin
    Posts
    1,103

    Default

    I can't find the PDP-1 picture and maybe it wasn't a Friden, but I did find this on Wiki, regarding PDP-1.

    "Offline devices are typically Friden Flexowriters that have been specially built to operate with the FIO-DEC character coding used by the PDP-1. Like the console typewriter, these are built around a typing mechanism that is mechanically the same as an IBM Electric typewriter.[22] However, Flexowriters are highly reliable and were often used for long unattended printing sessions. Flexowriters have electromechanical paper tape punches and readers which operate synchronously with the typewriter mechanism. Typing rates are about ten characters per second. A typical PDP-1 operating procedure is to output text to punched paper tape using the PDP-1's "high speed" (60-character-per-second) Teletype model BRPE punch, then to hand carry the tape to a Flexowriter for offline printing. "

    I also was talking to Phil Hayes at Bletchley Park via e-mail. He told me that the Friden shown was what they have now. Apparently the original Colossus used the IBM type writer. They could find an appropriate IBM machine that worked. Mike

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •