Dranetz 626 Repair/Redesign

[amspire]

I have had a dead Dranetz 626 Power Analyzer for years and I had started a repair about 2 years ago, and I decided it was time to complete the job. I have got it working, and my repairs are almost complete.



Built about 1984, it is still useful. Lots of plugin cards of all sorts are available and I have the 3 phase AC card, a low voltage and high voltage DC card, and a 8 channel even monitor card.

On the AC channel, it can monitor the voltage and detect problems including transients up to 4KV from 1uS in width. The modern analyzers from Dranetz can't do that. The front panel looks a bit daunting and unfriendly, but it is actually very easy to configure and use without a manual once you learn the logic. They have done a very good job with the interface.

For the following story, keep in mind this is a piece of quality gear. An optioned up 626 could have easily cost $8000 back in 1984. Enough money to buy a new car back then.

I  had a dead power supply in my 626. No idea if there were any other problems. It has an inbuilt 12V 1.6AH NiCd (that is a clue) battery and the 626 still had its original batteries.

I thought I would start with a question.  Here is the charger circuit:



What is wrong with this circuit? Yes this is the complete charger circuit - nothing is missing and the battery is permanently connected.

Here is a second clue - here is the state of this supply and a nearby LM339 voltage level monitor circuit on the plugin power supply board. I have removed the 2N5490 transistor, heatsink and LM339 IC. You can see bits of the tablecloth through the board. It was really thoroughly burned:




Another clue is the issue is not the reverse Base-Emitter breakdown of the 2N5490 transistor when the 15V +/- 0.05V supply is off and at 0V. I do not know how they got away with that one - special 2N5490's with a very high reverse base breakdown? I am also not sure of the power rating of the 10 ohm resistor but from the size, it may have been 1 to 2W and very good quality - perhaps wirewound. The circuit shows 1/4W but that is definitely wrong. According to the service manual, the charger gets the batteries to 14.3V where the current reduces to 100mA which the NiCds can take continuously. I cannot see how this is possible, but that is what the manual says. The thermistor is for temperature compensation for the temperature of the transistor to compensate for the change of Vbe as the transistor gets hotter. It is too far from the battery to be useful at battery temp compensation.

[SeanB]

Battery very shorted as well? This would always discharge via the poor transistor, ands it would go short circuit with time. Not helped by it being cooked as well, as it's gain drops with BE degradation. It probably finally died when the battery shorted out at last from being overcharged.

[NiHaoMike]

Looks like they originally designed it for a lead acid battery and then switched to NiCd without enough testing.

[amspire]

Battery very shorted as well? This would always discharge via the poor transistor, ands it would go short circuit with time. Not helped by it being cooked as well, as it's gain drops with BE degradation. It probably finally died when the battery shorted out at last from being overcharged.

Yes, half the cells were shorted, and with this brilliant charger circuit, the only current limit is the 10 ohm and 22 ohm resistor. If the battery voltage is 7V, then the 10 ohm resistor is dissipating 6.4W and so it gets to 500degC or something and cooks the PCB. It obviously triggered a few other parts to fail causing even more damage.

The charger circuit is designed so when the NiCds cells short - as they often do at the end of their life - the circuit board self destructs.

A total fail in my opinion.

The very weird thing is that most of the circuits in the 626 are massively over-designed. The PA-6002C DC monitor board I have only monitors a single DC channel.

It has an unbelievable total of 14 dual opamp ICs, 6 Harris analog switch ICs, 11 calibration pots and one variable capacitor. I don't have the circuit for this card, but the reason for the complexity is they have a differential input. They probably have separate channels for measuring the normal voltage (up to 600V), the transient voltage (up to 4KV) the transient period. They have to be able to capture sags and peaks so that probably means several peak detectors perhaps followed by sample and hold circuits. They have range switching, and they also have self calibration test circuitry that tests the A/D, peak detectors and transient width circuits. The point is they definitely haven't skimped.

So why do they have a battery charger with 5 components that destroys the power supply PCB?

Don't understand it, but at least it means I get a 626 Analyser very cheaply.

I decided to fix the board, and I wasn't going to put the same circuit back in. There is very little space, so I decided it was a good time to try out a PIC 12F675 8 pin smd micro that I had been meaning to try to make a smart charger. I like the idea of a charger that knows when the battery has had enough time on trickle charge and can drop down to a sustaining charge rate. Also it can detect undervoltage batteries and just stop trying to charge the battery.

I hadn't actually done a PIC micro project before - I have used other micros. I was keen to try out Hitech C again as I had used it for 8051 projects years ago and really liked it.

[amspire]

Here is my new improved charger circuit:



The board repair all worked out pretty well. The damaged PCB was cut out, and a new PCB epoxied in. As well as the new charger circuit, it also includes the LM339 voltage level detector circuit that was also destroyed. It all works, and the 626 now runs perfectly:




If you look carefully, there is a big clanger made by Dranetz. I took this photo before fixing it.

They put a 0.1uF 125V AC capacitor across the 240V AC mains input. The unit has obviously survived over 18 years with that capacitor. The parts list specifies a properly rated capacitor. It doesn't look like someone replaced it - it looks like that is how the supply was built.

Richard

[SeanB]

Original part probably was designed to survive and tested to 2.5 times rated voltage, so survived. probably was overengineered enough to survive 220V peaks and spikes as well.

[perdrix]

I've done my own charger design using an LT3086.  Current limited to just under 120mA and voltage limited to 14.4V, so it's a C/10 trickle charge into the original 1.2Ah NiCd battery pack, or a C/20 into a rebuilt pack using 2400mAH NiMH cells.

The reason I didn't use an LM317 constant current source is that there's not enough voltage headroom given the 15V supply and 14.4V being necessary to charge the battery pack (LM317 has 3V dropout).

I've done a PCB design and that is being made up by ITead Studio (board size 0.4" by 1.03").   I will have a number of spare PCBs when I get them and I've placed the PCB design in the public domain at ITead Studio.  More details once I've built the first one and confirmed its all OK.

The 100uF is a Case X (think case D but thicker) 25V tantalum, while C3 is a 1210 size 22uF, 25V X7R dielectric ceramic.  All remaining parts are 1206 except R2 which is 0805 (couldn't get the 0.1% part in 1206).

How do I get the image to appear inline???

[PA0PBZ]

How do I get the image to appear inline???

Attach the image, post, then get the properties of the expanded image and then edit and post the link in your original text between [ img ] and [ / img ]

[haef]

I too have the baked charger issue.  But dum question: Where on earth is the battery pack???  Either this is going to be a doh! moment, or someone removed the thing...

Would the original circuit work with a gel cell? 

I assume the 12v jack on the back is in only and does not provide a charge.

While I'm at it, what is a clanger?

[perdrix]

The battery pack is attached to a PCB which is in turn attached to the bulkead wall between the main body of the unit and the controls.

A clanger in british slang is an absurd or embarassing blunder (IOW huge mistake).   In context this seems correct.

Alternatively (not appropriate in this context (quote Wikiepedia)

Clangers is a British stop-motion animated children's television series of short stories about a family of murine creatures who live on, and inside, a small moon-like planet. They speak only in whistles, and eat green soup supplied by the Soup Dragon and blue string pudding. The programmes were originally broadcast by BBC1 between 1969 and 1972, followed by the first of three special episodes that was broadcast in 1974.

Dave

[perdrix]

I received the PCBs for this back from Itead Studio and made up one of them which I installed into the Dranetz power supply.

Here's a piccie of it installed:



It works just fine.

I have ten spare PCBs for anyone who wants to buy one, and I've open sourced the design at iTead studio under their SKU number IM120418001 if you'd like to order a set yourself.

I've put together a short document in the form of a PDF file showing how it all goes together and have attached it to this post.

Cheers
Dave

[perdrix]

A follow up on this.  The original NiCd pattery pack was shot.  So I ‘ve replaced it with four LiFePO4 cells in 26650 size.   The nominal capacity of these is 3200mAh.

Charge rate with the new charger will be C/27 (ish) and at that charge rate there's really no need to add a BMS.

Run time from fully charged to automatic power down was 1 hour and 40 minutes in my 626 with a 626-PA-6003 and two 626-PA-6002A installed.

Dave

[perdrix]

Sad postscript to this - I didn't use this for about 18 months or maybe 2 years, and when I came to use it next the LFP cells were dead.

Looks like two factors at work here:

• The battery pack cutoff voltage set by the resistive divider of R10 (124k) / R4 (46.4k) causes the relay to disconnect the battery when it's voltage drops below about 10V.  This is OK for NiCd cells but for a 4 cell LFP pack it is 2.5V per cell which is pretty much the damage limit!  It should be about 3V per LFP cell
  
• Even after disconnection the battery still supplies the memory board so it will continue to discharge below the damage level.

Lesson learnt don't use LiFePO4 cells where they can discharge below 2.5V

TIme to go buy some NiMHs

David