Ampeg BA600 - 115 bass amp combo, 6 months old, ROHS

[N_Cook]

Yet another owner forgoing replacement/repair within warranty. I imagine the
importer has no replacements in stock as the whole batch has the same
faults. Perhaps bass guitarists should use pneumatics and proper compressors
to shift air.
Mechanical rattle noise due to brass cylinder spacers used over the control
pot bushes , loose bush nuts, so rattle.
But more importantly
At moderate levels an electronic crackle is the main problem. Remove the
screws in the cab to release the amp and one caged nut falls into the amp
space due to cracked cage. And a nut ( now known) falls out of its cage in
the screened off ps+pa section. The PA although supposedly 600W uses only
TO220 ,(4 x TOP66 in the smps).
Amp and its casing 6Kg but 22x21x16 inch cab and 15 inch sp only 18Kg so
maybe part of the problem.
TO220 in the PA is 3 on pcb , 3 on one section of heatsink and 4 on another,
no fans. As slab SMR nearby are 22R I assume for the moment the o/p devices
are powerFET. Each bank of TO220 held down by cross bar and 2 screws.
This is my scale of screwdriver torque
1 light finger pressure around handle
2 strong finger pressure
3 fist
4 2 hands
5 wrench
6 impact driver
Testing the cross bar screws (they do have captive spring and plain washers)
torque measures about 2 , 1 and
then 1.5 and 0
The absent one either dropping out unseen as I knew a captive nut was loose
inside or still trapped under ps or pa.
So I assume silipads had compressed but these are 1.8mm thick porcelain,
even less compressible than mica. They were still in place surprisingle but
you could slide them with fingers and the TO220 nearest absent screw.
As distortion was crackle rather than push-pull failure distortion hopefully
solder failure , think I can see ring cracks with x30 microscope at awkward
angle, pushing around the loose TO220 legs not obviously moving at solder
points) rather than some thermal runaway devices.
So how to do an airframe type castellated nut plus lock-wire fix? for these
screws. 1mm wide pair of cross slots ground in the screw heads and st/st
wire looped and twisted in a slot and around the cross-bar?

 

[N_Cook]

Cannot find that 20mm long screw lurking anywhere and extremely unlikely to
jump through the grommeted lead ports into other section of amp and then
fall out. Straight on , top and bottom x30 viewing of all TO220 soldering
looks ok, surprisingly. Did find some nasty rework/fudge. 2 standard radial
legged 10uF 50V, C19 and C54 soldered to standard SM pads, no thru board
holes. Near it a SM cap, C61 measuring 15.2nF, had been soldered over
conformal coating with one end not soldered to trace, as coating under.
Owner has owned from new. 4x IRF450N power FET for 600W ?

 

[N_Cook]

Cannot find that 20mm long screw lurking anywhere and extremely unlikely to
jump through the grommeted lead ports into other section of amp and then
fall out. Straight on , top and bottom x30 viewing of all TO220 soldering
looks ok, surprisingly. Did find some nasty rework/fudge. 2 standard radial
legged 10uF 50V, C19 and C54 soldered to standard SM pads, no thru board
holes. Near it a SM cap, C61 measuring 15.2nF, had been soldered over
conformal coating with one end not soldered to trace, as coating under.
Owner has owned from new. 4x IRF450N power FET for 600W ?

Its prob'ly that class D junk....

+++++++

Valid point, it does use 2x Int Rect IRS20955S, each one driving a pair of
FETs and
there is an IR application note out there for those being used in Class D.
That
would explain the socking great inductor/choke well out of scale with all
the SM on that board. Will check that inductor is at the output and try and
locate a clock line , presumably back to the SMPS.

 

[N_Cook]

A generic problem with these will be mechanical noise from vibrating control
pots.
7 of the 9 were loose , in 6 months of non-commercial use. Not brass
cylinders over the bushes but are stop-ledged/ partial threaded leaving not
enough space for chassis thickness, nut and normal star washer so they used
thin plain washers. Luckily I have managed to fudge some thin star washers.

 

[N_Cook]

Downloaded the datasheet and that IRS is for pwm/D class operation. Input
,pin3, goes back via Q9 to 74C04 buffer so digital.
That large L , in a 53x53x24mm case is connected between the common points
of both FET pairs and across the speaker (with an ohm or so somewhere not
ascertained). That L1 measures 4K ohms , inductance not possible on LCR
meter as a yellow block HV polyprop? cap , label obscured by L1, is also
across it but measures 1.15uF apparently on LCR meter.

Speaker lead has a 4 pole connector for disassembly , 2 poles per wire, but
same size connector on the pcb only 1 pole each. Worse than that the commons
traces going to L1 are 3mm wide with solder run thickening for the current
but the 2 output traces going to the speaker are 1.5 mm wide and plain
tinned. So assuming 300 watt for a second into 4 ohm then about 9 amps.
From the Enumber in ul.com, it is 1 oz board and using
http://www.pcbco.com.au/tracecalc.html
for 60 mil track and 9 amps for plain copper then 100 deg C rise , don't
know how much the tinning coat would change things , say an unlikely
doubling to 2 oz in effect, then 32 deg C rise , unacceptable ?

 

[N_Cook]

Can you have the bridge of a bridge tied D-Class filter of form
insted of


--L--Sp--L--

as single L and speaker and C in parallel


|-----Sp-----|
----|------L-----|------
|-------C----|

L is inductor ,Sp speaker,C capacitor

 

[N_Cook]

I'd forgotten the double wound toroid inductor of 360uH in series either
side of the speaker.
What sort of value of inducance for L1 and its wire gauge/current carrying
requirement for 300 to 600W use? C may be about 1uF

 

[N_Cook]

I removed the filter cap to measure the inductance of L1 . Cap marked 1uF,
250V.
This L1 weighs about 170 gm, estimating weight of remainder components and
pcb from total of 290 gm.
Through a hole I can see some wire about 1mm diameter.
Using 1KHz RLC meter, L1 measures -2.2H on H scales and -140mH on mH scales,
don't know what that means , will try and find a manual for an Avo B183.
Maybe faulty, but measuring plain inductors through the scales is positive
and consistent across neighbouring scales.
1KHz ac resistance is 230R, DC resistance is 4K so perhaps what is labelled
as L1 is R and L in series.
Perhaps R of 4K and L of .6H , to give 230R at 1KHz, but if wholly 1mm wire
I cannot see how 600mH with 170gm of 1mm wire, strange core material?

 

[N_Cook]

It almost seems perverse, 0.4 amps of mains current draw (at 240V) and 25
watts of 400Hz sine into 4R dummy load and a thermometer laid over the tabs
of the 4x TO220 output device, rises only 8 deg C over ambient.
And that with amp laying on bench so no convection over the external
heatsink vanes, no wonder no fan needed.

Previous D class I've played with were Stagepas , not possible to monitor
their operation without making up umbilicals to operate PA out of the little
box. And as separate clock maybe "RF" problems doing that

 

[N_Cook]

Do Ampeg make bass amps? There will be loads of returns of this model.
Any large lump inside a bass amp is likely to vibrate if inadequately
secured. That 2 x 2 x 1 inch lump is only secured by its screening can at 2
pointsplus inductor ends soldered. That was the main problem , proper solder
would be hard pushed but PbF only lasts 6 months like this. By a daisy
chaining cable ties between made anchor point at the ps , round the L1 can
and out through the cable hole and tightened after the ps+pa screening case
is back over.
Also PbF failing on the speaker socket, would not have been so bad if they'd
used the 4 pole connection like farther down the speaker line. Plus the
absence of star washers on the pots. Ignoring the absent h/s bolt.