tep392 Posted June 9, 2022 Share Posted June 9, 2022 I have a 1050 that I modded a while back with the super archiver/bitwriter replica. I use it frequently and it has been very reliable. I decided to replace the original linear 12V and 5V regulators with switching regulators, and now it won't work reliably with any of my disks. I can read some disks, but it's hit or miss. Most bootable disks get boot errors and even if they start to boot, they may not finish. The exception is that I can format a disk on the drive and that disk will work perfectly. The newly formatted disk will even work in my other 1050 without problems. I checked the voltages at TP13 and TP14 and get 5.4V and 12.9V respectively. Any thoughts on what might be happening? I tried checking the speed in case it was off, but the test only works if I use a disk formated on the drive. Then I get 289 rpm. Quote Link to comment Share on other sites More sharing options...
TGB1718 Posted June 9, 2022 Share Posted June 9, 2022 Those voltages are out of spec for a 7805 and 7812 should be in the range 4.8 to 5.2 VDC and 11.5 to 12.5 VDC. It may also be they are very noisy depending on what frequency they are running at. It may be possible to reduce the noise by adding further smoothing capacitors across the outputs (if you have room) Quote Link to comment Share on other sites More sharing options...
tep392 Posted June 9, 2022 Author Share Posted June 9, 2022 I bought from from EzSBC. I've sent an e-mail to ask about the voltages being high. Do you think that could cause my issue? The switching frequency is 500kHz. I'll put my scope on them later and see how much noise there may be. Quote Link to comment Share on other sites More sharing options...
tep392 Posted June 9, 2022 Author Share Posted June 9, 2022 I just heard back from EzSBC and they say every part is tested to be within 100mV under load. I'll have to take a look with my scope to see what is going on. Quote Link to comment Share on other sites More sharing options...
tep392 Posted June 9, 2022 Author Share Posted June 9, 2022 I guess my multimeter is bad. I checked with the scope and the voltage is well within spec. There is about +/-20mV of 500kHz ripple. Quote Link to comment Share on other sites More sharing options...
tep392 Posted June 9, 2022 Author Share Posted June 9, 2022 I made an interesting discovery. I put a probe on TP5, which is the output of the read head amplifier. I had to lift the mech out of the drive about an inch to make room for the probe. It started working again! I then removed the probe and put the mech back in place, and got boot errors. While getting the errors, I lifted the mech back out and it started reading the disk! I spent some time playing around with the wires and connectors, thinking there could be a bad connection, but the only way to get it to work is to lift the mech out. I don't see any contact between the bottom of the mech and the circuit board components. Getting closer. Quote Link to comment Share on other sites More sharing options...
tep392 Posted June 10, 2022 Author Share Posted June 10, 2022 I was looking at my other 1050, which has the original linear regulators, and the signal at TP5 is clean. Looking at the same signal on the problem 1050, there is very noticeable jitter in the signal. As I lift the mech away from the PC board, the jitter goes away and the signal gets clean and it starts to read. I'm still not sure why I can format a disk on the drive and it's readable, but my other disks are not. Hmm. Quote Link to comment Share on other sites More sharing options...
_The Doctor__ Posted June 10, 2022 Share Posted June 10, 2022 put the linears back in and see if it works... 2 Quote Link to comment Share on other sites More sharing options...
tep392 Posted June 11, 2022 Author Share Posted June 11, 2022 1 hour ago, _The Doctor__ said: put the linears back in and see if it works... That was going to be my next step after I exausted everything else. I did find that the ground wire that clips on the transport was broken, so I fixed that but no change. I swapped the 12V linear regulator in first. My thinking was that it supplies the analog amplifiers in the read circuit, and the digital stuff seemed to be functioning normally. There was no change. I put the 5V regulator back in and now it's working fine. I don't understand why lifting the transport out resolves the issue. I'm going to experiment with some filtering on a breadboard and see if I can reduce the ripple significantly before trying them out again. 1 Quote Link to comment Share on other sites More sharing options...
cwilbar Posted June 11, 2022 Share Posted June 11, 2022 3 hours ago, tep392 said: That was going to be my next step after I exausted everything else. I did find that the ground wire that clips on the transport was broken, so I fixed that but no change. I swapped the 12V linear regulator in first. My thinking was that it supplies the analog amplifiers in the read circuit, and the digital stuff seemed to be functioning normally. There was no change. I put the 5V regulator back in and now it's working fine. I don't understand why lifting the transport out resolves the issue. I'm going to experiment with some filtering on a breadboard and see if I can reduce the ripple significantly before trying them out again. What regulators did you use, Traco ? or something else ? wonder if the coils on the head are picking up this high frequency ? I believe I've read about Traco power being used on 1050s, maybe different DC to DC converters operate on different frequencies and/or have better internal filtering ? Quote Link to comment Share on other sites More sharing options...
1050 Posted June 11, 2022 Share Posted June 11, 2022 1 hour ago, cwilbar said: What regulators did you use https://www.ezsbc.com/product/psu3-5/ https://www.ezsbc.com/product/psu3-12/ Quote Link to comment Share on other sites More sharing options...
BillC Posted June 11, 2022 Share Posted June 11, 2022 7 hours ago, tep392 said: That was going to be my next step after I exausted everything else. I did find that the ground wire that clips on the transport was broken, so I fixed that but no change. I swapped the 12V linear regulator in first. My thinking was that it supplies the analog amplifiers in the read circuit, and the digital stuff seemed to be functioning normally. There was no change. I put the 5V regulator back in and now it's working fine. I don't understand why lifting the transport out resolves the issue. I'm going to experiment with some filtering on a breadboard and see if I can reduce the ripple significantly before trying them out again. It doesn't take that much AC ripple to cause an issue with DC electonics. When I built a custom +5V power supply with crowbar protection from scratch over 30 years ago my 800XL wouldn't power up with it initially, in spite of having the correct DC voltage. I found the output also had a small amount of AC ripple, changing one of the output capacitors reduced the ripple enough to resolve the problem. Quote Link to comment Share on other sites More sharing options...
TGB1718 Posted June 11, 2022 Share Posted June 11, 2022 (edited) On 6/10/2022 at 12:18 AM, tep392 said: There is about +/-20mV of 500kHz ripple. The regulator frequency of 500KHz = second harmonic frequency of 1 MHz exactly the same as the clock frequency in the 1050, I believe the 4 MHz clock in the 1050 is divided by 4 to produce this, so if the regulator is particularly noisy the it going to interfere with the main clock frequency. As I suggested earlier, try increasing the smoothing capacitor on the 5V line It may be the 12V line is not so noisy as there's a 100uH inductor in series with the output. Edited June 11, 2022 by TGB1718 2 Quote Link to comment Share on other sites More sharing options...
tep392 Posted June 11, 2022 Author Share Posted June 11, 2022 11 hours ago, cwilbar said: What regulators did you use, Traco ? or something else ? wonder if the coils on the head are picking up this high frequency ? I believe I've read about Traco power being used on 1050s, maybe different DC to DC converters operate on different frequencies and/or have better internal filtering ? EzSBC PSU5a and PSU3-12. I think the head is picking up noise, since it start working when I move the transport about an inch above the circuit board. I'm going to try some filtering caps and find a way to rework the shield so I can fit it over the bitwriter replica board. Quote Link to comment Share on other sites More sharing options...
+Peter Rabitt Posted June 11, 2022 Share Posted June 11, 2022 OK you may have the answer and I just may have found the answer to the xf551 that was working great until I added the 5v switching reg. Then it stopped working but the heat sync was cool to touch. Now I am going to go back and look at the ripple and see if it has problems. If you have a scope with 1x probe you can use it as an antenna and just put it near the reg. and see if the "noise" increases or not. Do not use the ground probe to ground just use the probe as a antenna. O,yes .001uf or .01uf caps on the output side would help.. The only problem I see is the "ripple/interference" maybe RF then you may have to remove the reg.s all together... Peter Quote Link to comment Share on other sites More sharing options...
tep392 Posted June 11, 2022 Author Share Posted June 11, 2022 Here are a couple thoughts from the EzSBC designer. These regulators have filter caps built-in and the external electrolytics, C69 and C70 could be too much and possibly cause instability in the regulator control circuit. He suggested removing them. The regulators also emit some magnetic fields due to the inductors, which could be getting picked up by the read head. The fields are directional, so changing the orientation of the regulators could help. Quote Link to comment Share on other sites More sharing options...
tep392 Posted June 12, 2022 Author Share Posted June 12, 2022 I had some success today! I installed the EzSBC regulators in my other 1050 which is a Tandon unit with a Mini-speedy installed. I removed the C69 and C70 caps which aren't needed and could cause stability issues with the regulator. I also installed the RF shield which I removed when the speedy was installed. It has just enough clearance with the Mini-speedy. The drive works perfectly fine! I tested it with single and double density disks and commercial disks which are all working. 1 Quote Link to comment Share on other sites More sharing options...
1050 Posted June 12, 2022 Share Posted June 12, 2022 My input would be to place some ferrite beads on the output leg and glue them on with super glue. Then perhaps place a couple .33 uf chip bypass capacitors to shunt RF signals on both the power to ground planes and output to ground plane. Seems to be areas quite available if you'll scrape a bit with a sharp knife. Don't think this one is radiating EMF too much, that's pure operating loss and while it's always possible, the effective field is well constrained in the inductor design from day one and I don't expect it's being overpowered to any significant degree to cause uncaptured magnetic lines of force to heat my coffee up with. Selecting Decoupling Capacitors for Atmel’s PLDs http://ww1.microchip.com/downloads/en/Appnotes/DOC0484.PDF A decoupling capacitor is also known as a standard bypass cap when you shut off the aggrandizement double speak filter. Why people insist on the buffalo approach as a way to explain something pretty basic is a curse of the ego run amok. At any rate, our Atari are using the glass encapsulated (bodgy/dodgy) .1uf standard unit, commonly marked as 104 (10 and four zeros for 100,000 pico farads) very often and they always were inadequate in the first place. They took a perfectly good chip capacitor and put axial leads on it and then rolled it up inside a glass case such that it looks very much like an old school diode. Quote Link to comment Share on other sites More sharing options...
+Peter Rabitt Posted June 13, 2022 Share Posted June 13, 2022 Question to help me fix/repair some of the floppy drives I have.. This may not have anything to do with the switching reg. but to me it does... I am looking into the xf551 I have that I changed the 5.0v reg to a sw type and lost the working drive... Because I have a few drives I would like to know what drives are compatible... 1050 w/no changes (stock) 1050 w/doubler? does sd/dd. 810 w/doubler? does sd/dd Randa w/no mod. sd/dd Ingus w/65k mem. upgrade Ingus stock. xf551 sd/dd/ss/ds (working on this one) Now if I format a disk with any of the above will the others read them?? Trying to see which ones work and if I have fixed some of the problems.. Is their any one that will work with all of them... I have some drives that try to read a formatted disk but stop.. Some try to format a disk and the first time thru its OK but but the sec. try VER. stops at ?39?then back up to ?38? and stops... Format on the 810 and try to read it on the 1050 does not work, etc... Peter Quote Link to comment Share on other sites More sharing options...
TGB1718 Posted June 13, 2022 Share Posted June 13, 2022 Single density format on any of those drives should work with all of them. Enhanced density obviously on the 1050's (don't know if the others should be able to read enhanced except the 810 which can't) Double density should also work across the double density drives only. And of course only the xf551 can read double sided The obvious things if formatting is inconsistent, clean the heads, check the floppy's make sure the surface is nice and shiny, no tramlines made by the head. If after all that you still have problems between drives, than it sounds like you need an alignment disk, chances of getting hold of one of those if like getting rocking horse s**t If you have a commercial game or utility disk, you could use that to align the drives. Quote Link to comment Share on other sites More sharing options...
+Peter Rabitt Posted June 13, 2022 Share Posted June 13, 2022 Yea Thanks for the info.. I did think they should read the same disk in the proper density but have been having trouble going between the different drives... I guess it time to check out the calibration of the drives... I think I will also change beck to the 5.0 linear reg. and take out the 5.0 sw reg. I will live with the heat on the heat sync. I think it is the pain...HAHAHA Again thanks for the info... Peter Quote Link to comment Share on other sites More sharing options...
_The Doctor__ Posted June 14, 2022 Share Posted June 14, 2022 last point of note, the RPM's need to be correct for each drive... otherwise they can fail to format and verify causing retries. Quote Link to comment Share on other sites More sharing options...
+Peter Rabitt Posted June 14, 2022 Share Posted June 14, 2022 TK'S DOC!!! That would be part of the calibration of the drive... In my world.. Just out of making sure what is the single side, single density RPM for the above drives... I know the xf551 is 300? rpm but are all the others 288 +- 2 rpm or so??? peter Quote Link to comment Share on other sites More sharing options...
TGB1718 Posted June 14, 2022 Share Posted June 14, 2022 33 minutes ago, Peter Rabitt said: I know the xf551 is 300? rpm but are all the others 288 +- 2 rpm or so??? Remember any modified drives will not work with the Atari Diagnostics disk, you will get strange readings, I think some people use a strobe to determine the speed of those drives, not done it myself, but sure someone will explain how they do it. Quote Link to comment Share on other sites More sharing options...
_The Doctor__ Posted June 14, 2022 Share Posted June 14, 2022 the strobe works the same as on record players the pattern blocks will solidify and float slowly forward or backwards until you dial it in... at which point the blocks will appear pretty much stationary with a slight jitter every now and again during operation. Quote Link to comment Share on other sites More sharing options...
Like us on Facebook
Follow us on Twitter
AtariAge on Instagram
Watch us on YouTube
Chat on Discord
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.