DECmate II Power Supply Service

The power supply sits at the rear of the DECmate II system unit, behind the RX50 drive (DEC, DECmate II Owner's Manual, Figure 1-8)

The DECmate II runs from an H7842-A switching power supply, the same supply DEC fitted to the Rainbow 100. DEC published no service manual for the H7842 itself, but chapter 7 of the Rainbow 100 Technical Manual (EK-PC100-TM-001) describes the circuit and gives the full input and output specification, and Tony Duell drew a schematic with component values from a working unit. The DECmate II owner's manual treats the supply as a module to be replaced whole.

Warning: the H7842 rectifies the mains to about 325 V DC. DEC's instruction is to switch off, wait five minutes before removing the cover, and confirm with a voltmeter that no high voltage remains before handling any component.

Identification and connections

DEC's DECmate II print set lists the H7842-A in the BA25 system unit and describes it as a 138 W small-system supply. The owner's manual calls it a 135 W supply.

• Input: IEC mains connector, recessed circuit breaker and a slide switch for 115 V or 230 V on the back of the unit.
• System board: one 13-pin flat cable, detachable at both ends.
• RX50 drives: two 4-pin connectors fixed to the supply.
• Power switch and fan: two cables from the front assembly plug into the supply.

The supply is released by a slide tab at its end.

On the Rainbow 100, DEC gives this pin-out for the 13-pin connector (J8) on the system module. The DECmate II manuals do not print it.

H7842 to system module, 13-pin connector (Rainbow 100, J8)

Pin	Signal
1	ACOK H (AC voltage OK)
2	VBIAS
3	Key (no pin)
4	−12 V
5, 6	+12 V
7, 8, 9	+5 V
10, 11, 12, 13	Ground

Specification

H7842 outputs (Rainbow 100 Technical Manual, section 7.5.2)

	+5 V output	+12 V output	−12 V output
Nominal and load range	+5.1 V, 2.5–11.5 A	+12.2 V, 0.6–6.7 A	−12 V, 0–0.15 A
Total tolerance	±6%	±6%	±7%
Ripple and noise	50 mV p-p	75 mV p-p	120 mV p-p
Overcurrent trip, minimum to maximum	10.7–14.5 A	6.8–10.0 A	0.4–1.5 A
Short-circuit current, maximum	8.0 A	4.0 A	3.0 A
Overvoltage trip, minimum	5.80 V	–	−13.0 V
Overvoltage, absolute maximum output	7.0 V	–	−15.0 V

The input is single-phase 90–128 V rms or 180–256 V rms, at up to 3.0 A (115 V) or 2.0 A (230 V), 226 VA apparent and 138 W maximum including 23 W for the fan. The supply holds its outputs for at least 20 ms through a mains dropout at low line and full load. The summary in chapter 1 of the same manual gives the +12 V output as +12.1 V and the upper input range as 174–256 V. The DECmate II owner's manual also gives 174–256 V.

Checking the outputs

Measure on the 13-pin connector or the drive connectors with the supply loaded by the system. Using the tolerances above:

Acceptable output range (from DEC's total tolerance)

Output	Nominal	Acceptable range
+5 V	5.1 V	4.79–5.41 V
+12 V	12.2 V	11.47–12.93 V
−12 V	−12 V	−11.16 to −12.84 V

How the supply works

DEC's description, from the Rainbow 100 Technical Manual:

• One side of the mains passes the EMI filter, a 4 A circuit breaker and the front-panel power switch to the fan, the start-up transformer T1 and the bridge rectifier D1. The 115/230 V switch turns D1 into a voltage doubler on 115 V, so the bus is about 325 V DC either way.
• A fuse beneath the EMI filter protects the fan if the wrong mains voltage is applied. With 230 V applied and the switch at 115 V, DEC warns that internal components will be damaged; with 115 V applied and the switch at 230 V, the fan runs slowly and the computer does not work.
• T1 also supplies V BIAS, the DC supply for the control module.
• The switching transistor Q1 drives the primary of transformer T3. A pulse-width modulator on the control module drives Q1 through the switch driver Q4 and transformer T2.
• T3 has three secondaries. The +5.1 V and +12 V outputs are forward converters; the −12 V output is a flyback winding shunt-regulated by Q2, Q3 and zener D15.
• R9 and R14 sense output current. The control module also watches for overvoltage, low bias voltage at start-up (lockout), transformer saturation and an excessive volt-time product, and can shut the PWM down.
• An overvoltage on +5 V or −12 V fires the crowbar D10 across +5.1 V, which forces the supply into overcurrent shutdown. The crowbar can only be cleared by resetting the supply, that is, switching off and on.

Capacitors

Tony Duell's schematic gives these electrolytic values.

H7842 electrolytic capacitors on Tony Duell's schematic

Position	Value	Duell sheet
Mains reservoir, two in series across the doubler	820 µF (×2)	1
Start-up supply, raw 12 V before the 7812 regulator	2200 µF	1
V start, after the 7812	10 µF and 2.2 µF	1
+12 V output, before the filter choke	3200 µF	3
+12 V output, after the choke	47 µF	3
+5 V output, before the filter choke	5600 µF	3
+5 V output, after the choke	47 µF	3
−12 V winding	560 µF	3
−12 V after the regulator	330 µF and 47 µF	3
−12 V regulator	2 µF	3

Duell's drawing gives no voltage ratings. A list posted to the cctech mailing list in 2019, quoting a VCF forum post, gives ratings for the same set: 2× 3200 µF 16 V, 1× 2200 µF 35 V, 2× 820 µF 250 V, 1× 560 µF 20 V, 1× 330 µF 20 V, 3× 47 µF 16 V, 1× 10 µF 35 V and 1× 2 µF 25 V. That list has two 3200 µF parts where Duell's sheet 3 shows 3200 µF on +12 V and 5600 µF on +5 V. Supplies may differ between revisions, so read the value and rating printed on each capacitor before ordering.

The 820 µF reservoirs sit across the 325 V bus. Fit parts of at least the original voltage rating and check the board's printed polarity.

Repair record

Rob Jarratt's account of repairing a failed H7842 from a Rainbow is the most detailed public record of this supply's faults. In order:

1. The machine switched itself off with a burning smell and the circuit breaker tripped. Re-applying power gave a bang: the main switching transistor, a BUV48A on Duell's sheet 2, had disintegrated.
2. The diode from the transistor's base to the drive transformer, in parallel with a 2.7 Ω resistor, was short and broke when lifted. It was replaced with a UF4007 on advice from the cctalk list.
3. The 7812 start-up regulator had also failed.
4. With 15 V from a bench supply into the 7812 input, a working control board drew about 87 mA and the PWM drove the chopper. The faulty one drew about 140 mA, its 5.1 V reference read 9.75 V and the PWM gave no output, so the PWM was replaced with a UC3527AN.
5. The PWM's shutdown input was still asserted. The LM339 comparator E3 was replaced. Jarratt reports that Duell's schematic appears to swap the labels of two sections of E3 (E3c and E3d) in the current-sense circuit.
6. An overcurrent indication on −12 V during bench tests later disappeared on its own.
7. The start-up regulator input was only about 3.5 V. The mains bridge rectifier had two pins shorted, probably when the switching transistor failed. With the rectifier and the start-up diodes removed, the start-up transformer gave 24 V AC. A new 600 V bridge restored the supply.

Jarratt tested with two 100 W filament bulbs in series with the mains and a dummy load on +5 V and +12 V. On a working supply the bulbs pulse and stay dim; on the faulty one they lit brightly and the outputs stayed at zero.

In a second Rainbow, an H7842-A stopped with a pop when the mains cable was plugged in, before the switch was turned on. The Corcom F2987A power entry module measured short across its input. Its mounting is not standard; Jarratt opened the soldered can, found the filter potted in silicone, and in the end fitted a Schurter 5121 snap-in module on a steel adaptor plate.

Related pages

• DECmate II
• DECmate II General Maintenance
• DECmate II Troubleshooting Guide
• DECmate II Documentation
• Capacitor Failure Symptoms
• Recommended Tools