------------------------------------------------------------------------

    Main directory    QL-Recorder    Programming    Colours & Notes    Literature    Links    What's new?   

------------------------------------------------------------------------

Contributed by Robert Burgess:

Fix for an Emulator Ultra losing its display after several minutes,
bad power supply ground plane identified as underlying cause

------------------------------------------------------------------------

To start with a legal statement: Registered trademarks are the properties of their owners; I do not imply that E-mu Ultras typically should need the fix described herein; the provided information is intended only for scientific or amateur use; please keep in mind that trying to service your E-mu on your own may void your warranty; please keep in mind that at least the system power supply and the backlight power supply on the motherboard may produce dangerous voltages. I do not accept any liabilities for the results of your own action, even if you might claim that I should have provided some inspiration. If you continue to read, this is on your own risk, and if you think that the contents of this page might violate any law in your country or any right or anything else that they should not violate, please be informed that I do not allow you to read any further. If you should know that there might be a legal problem with this page at all, please do not hesitate to send me a free (!) e-mail with further information in order to enable me to comply with any necessary regulations. And finally, if somebody should open up a business doing this procedure on many Ultras or if some hardware designers should not have known this solution or its relevance to specific existing or future designs (at E-mu or elsewhere), and thus this page should be a contribution to someone make a huge profit - please feel free to share a bit of it and support my projects described elsewhere on this web site. :-)

Please note: Switching power supplies carry extremely dangerous voltages, even after they have been removed from the power outlet! If you are not specifically trained, do not attempt any work in this place. Leave the diagnosis and repair of the problem described below to some qualified person!

------------------------------------------------------------------------

Robert Burgess from Tektrax AV Analytical (US), who sent me this fix, may be reached at: ae1tektraxplease_place_the_usual_symbol_herehotmail.com

As of 4/18/05 Monday

E-MU E-5000 Ultra / Digital Sampler / Tech posting

I'm sure the techs familiar with E-MU products knew of this problem and cure.

Finding no information on the net, I posted this.

Complaint

(Losing display after few minutes of being on)

Owner said it's been this way over a year. Just after powering up, within 5 min to an hour, the display lost all information and the back light at the same time. Test confirmed the complaint with an added characteristic. Sometimes the display would not reappear after repeated startups. Turning it off again, allowing adequate capacitive discharge time, upon next power up the display would reappear normally and first problem would repeat.

Diagnosis

On the road available tools

Dual Trace Scope / Digital Multi-Meter / Giga-Hertz Freq Counter / Computer with Net Access

No Schematics; no beneficial Internet search results; no documented help via E-MU web sites.

To make a very long story short, +5 volt is supplied to the display board, via one multi connector IDE style cable from the motherboard to the display, and a separate pair, +5 v (plus) ground from the motherboard to the back light.

Tattle-Tails

Three megahertz could always be seen unchanged at the display's clock crystal, with 2.4 volts +/- both sides of the crystal; at least at the times they were measured.

Monitoring 5 volts at the BLASTER ID Jumper Pin Jack W4 (had no jumper), measurement started at 4.89 volts at power up and slowly dropped to 4.63 within 10 minutes, at which point the problem would occur. Several repeats later, the next time, both display information and the back light weren't seen. The 5 volt line powered up at 4.62 volts. After this, a lengthy "power off" letting the capacitors discharge, the original problem was recreated.

Discovery

To check the supply, it was un-mounted and exposed with connections to the mother board, then using two alligator clip leads, provided grounding of the supply to the chassis (two separate supply board ground prints).

Before powering up, the 5 volt area of the board was identified showing a Shockley Device providing the +5 volt rail. At power up, the Shockley's output was 4.91 volts. It's following load resistor's output to the motherboard was 4.89 volts. The display gets it's 5 volt line from the motherboard which measured the same as at jack W4. The voltage remained unchanged for three hours without a fault, twice tested.

Next Best Guess

Assuming that the supply was being interfered with before the unit was disassembled, previous supply board grounding would be a key suspect. Visually checking the supply board's prints (single sided), no faults could be seen via magnification, therefore, ground points via screws is next.

Solution

Mod

In order to eliminate grounding problems via screw mounting, with the changed ground terminating to be chassis ground for the AC power line, a braid of copper (solder wick) with clear heat shrink tubing was used to create a ground strap across the supply's board from ground corner to ground corner, starting the strap at the farthest grounding point in relation to termination, connecting to the next board corner print without breaking the braid, continuing to chassis ground, and piggyback soldering the braid to the AC's ground line.

Using a Dremel and a small grinding wafer, the excess solder was removed from the two new grounds for the supply, allowing the screws to flush secure the board to its stand offs. It's well enough to grind the solder completely away from the board's "screw hole" ground prints since the braid now does the ground connections, with or without the screws.

Retest Setup
Retest

Final

Bad Power Supply ground plain caused by grounding screws loosening over time.

Could re-tightening the screws fix it? Sure- but still effected by repeated motion and environmentals.

Why not use standard insulated wiring in place of the braid? At the time, braid was all I could find.

Even at my shop, many times I've used braid to create better grounds and even voltage bussing:

Now, 48 hours and it's still on with voltage measurements holding steady.

Fixed.

Once again: Warning! Danger!

While the little transformer and the little switching power supply for the display backlight may give you unpleasant or deadly shocks, the large switching power supply for the whole machine may do so even more easily. So please obey my warnings, and leave diagnosing and repairing this problem to a qualified person.

Please note: Any power supply, when connected to the mains outlet, has wires that carry live voltage. Touching them accidentally, or unintentionally connecting them to a metal case, may be hazardous. Moreover, switching power supplies may carry extremely dangerous voltages, even after they have been removed from the power outlet! If you are not specifically trained, do not attempt any work in this place. Leave the diagnosis and repair of the problem described above to some qualified person! This remark is no kidding, switching power supplies may generate potentials of more than 1000 V at high frequencies - touching the wrong place may hurt you very unpleasantly, or even kill you.

Related pages:

------------------------------------------------------------------------

    Main directory    QL-Recorder    Programming    Colours & Notes    Literature    Links    What's new?   

------------------------------------------------------------------------

Woopie This Web page was prepared by myself,
and so were all included graphic elements.
© 14.11.2005, 23.03.2009 Dr. med. Jörg M. Sigle, Kunstvolle EDV & Elektronik

------------------------------------------------------------------------