Testing and using a few cheap Chinese
items
*****
A Digital Voltmeter for valve equipments
|
 |
It pays to read and fully understand the various
figures quoted with all these Chinese items. In this case the
spec of the on-board regulator quotes a 26 volt max rating which
is below the figures 30 and 40 volts in the advertisment. |
 |
|
|
A neat way of monitoring voltages in a high
voltage power supply or even fitted in a valved equipment is
to use a commercial display like the one shown above (I bought
a pack of five for £6.79 via Amazon). The difficulty at
present is that all those available tend to be designed for relatively
low voltages or driven from the voltage one is measuring which
means one must choose what is termed a "3-wire" display
module where separate wires are used for powering the module
and for monitoring a different voltage. Of course, to overcome
the low voltage limitation, a display can be used in a similar
way to using an analogue meter by fitting a limiting component
such as, in this case, a potential divider at the voltage measurement
input wire (yellow in the above example). Some displays may have
a limited range of values for the leftmost digit so one must
bear this in mind when designing the potential divider in order
to display a full range of leftmost numbers. |
|
 |
The particular device pictured needs an operating supply
of 4 to 30 volts and can display 0-100 volts. Say we'd like to
display the output of an HT power supply capable of 500 volts,
the display will need to show 50.0 volts. A potential divider
will need to be designed to convert 500 volts to 50.0 volts.
The first design requirement is to determine the monitor current
taken by the display module when its driven by your preferred
supply voltage because this will have a bearing on the potential
divider resistors.
The supplier quotes "less than 20mA supply voltage"
and I assume the lower the supply the less heat dissipation in
the module. Bearing in mind of course that some HT supplies have
their negative feed isolated from ground (=chassis) in order
to provide a bias supply (eg the R1155) so a little thought might
have to be given to the operating voltage for powering the display.
Reverse engineering devices like these is tricky because
chip markings are often erased (as in this case) so one must
carry out bench tests to decide on the optimum resistor values
for the potential divider. Having got the display to read 50.0
volts (or 5.00 volts) for our 500 volt input the next step is
to remove the decimal point so that the reading is a true 500
rather than 50.0. This might be straightforward, as we need no
decimal point for a three digit HT supply and, the DP pin is
accessible at Pin 3 of the display chip.
Left is a drawing of a "3631" (3631AW, LSH3631AUB,
LSH3631AUR, LSH3632AUY or YU-33611BG etc) |
 |
 |
|
|
With zero current fed into the monitor point, divider
resistors can be 220Kohm and 22Kohm (converting 500v to 50.0v)
or 220K and 2.2K if we need to convert 500v to 5v. Without experimenting
I can't say which divider is needed in order to correctly activate
all three digits (ie 5.00v = 500v rather than 50.0v = 500v).
When say 1mA is absorbed by the monitor point this current is
supplied via the 220Kohm resistor so the resistor value needs
compensating for this. One feature of this application is that
the accuracy needs to be far less than say over the full range
of 0-500 volts as we're really only interested in displaying
a limited range about a nominal value (eg 325v). Some trial and
error is probably the best approach for determining the the actual
value of the "220K" resistor or its mate the "22K"
resistor for our output voltage design range.
There's an SOT89 input chip carrying either the marking
"7133-1" (below) or "9956A" on a second example,
which I assume to be equivalent and which I assume will be the
power regulator for the display circuitry. A 16-pin chip (with
its markings erased) will be the display driver and basic voltmeter.
The 3631 display chip carries merely a set of LEDs whose cathodes
are grounded, meaning that the unwanted decimal point pin (Pin
3) can be disabled by altering its voltage via a grounded resistor
to the LED cut-off point. For a permanent HT voltmeter cutting
the DP pin, or removing solder from around the pin, might be
the sensible option. |
|
 |
|
Testing the display using a variable voltage PSU proved
it worked from around 4.6 volts up to 30 volts. I didn't increase
this beyond 30 volts as the spec indicates this as the max working
voltage (despite the 24 volt limit in the HT7133 spec?). Connecting
this supply voltage to the yellow wire proved that the total
current consumption was around 13 to 17mA. One slightly puzzling
fact was inputting 9 volts drew 13/14mA but above this increased
the current to 17mA but then the penny dropped... the illuminated
digits of 9.0 volts increased to say 12.0 volts and the additional
draw from the end digit was responsible for the 3mA current increase.
Measuring the current draw by the yellow wire at 24 volts was
a little over 100uA. so the potential divider resistors can now
be worked out. Turning on my variable voltage HT supply and setting
this at 85 volts correctly displayed this figure and slowly cranking
this up revealed a max reading of 99.9 volts. This means that
the leftmost figure isn't limited so a ten to one divider can
be used. Setting the HT resistor to 220Kohm will draw (500-50)volts
over 220K = 2.045mA with the ground resistor as 50 volts over
(2.045-0.1)mA = 27Kohm. Wattagewise 220kohm @ 2W and 27K @ 0.5W
should be fine. Below with a supply voltage of 12 volts and resistors
marked 220K and 27K actually measured 217K and 26.7K. |
|
True voltage |
Indicated voltage |
|
503.0 |
47.2 |
|
400.0 |
37.5 |
|
304.2 |
28.3 |
|
200.8 |
18.6 |
|
151.2 |
13.9 |
|
100.0 |
9.1 |
|
Then I added a 500K pot across a 39K resistor
in place of the 27K resistor and adjusted the pot for more accurate
readings (below), ending with a measured resistance of 24K which
seems odd at first sight. |
|
True voltage |
Indicated voltage |
|
401.0 |
40.5 |
|
298.0 |
29.9 |
|
201.0 |
20.1 |
|
149.6 |
14.9 |
|
Next I'll need to get rid of the decimal point.
This turned out to be harder than I'd imagined. Grounding Pin
3 of the display resulted in excessive supply current, as did
connecting Pin 3 to the supply voltage. I found the optimum settings
were to reduce the supply voltage to 5 volts and connect Pin3
via 50 ohms to the supply voltage. This had the effect of blanking
the decimal point at the expense of an increased supply current
from a nominal 15mA to about 27mA. An option would be to either
cut the connection to Pin3 or perhaps apply a dot of black paint.
Isolating Pin3 by removing the solder might also be possible? |
|
|
|
|
I looked at the various options; none of which
was attractive so decided to trace the connection to the DP.
It seemed to me that the likely source was a driver and checking
the pins to the adjacent chip, U1 (the one with its identification
erased) proved its pins were connected directly to the display
(ie Pin 8 you can clearly see goes to display Pin 1, Pin 7 buzzed
out to display Pin 2 and Pin 6 to display Pin 3). I carefully
lifted Pin 6 from its pad (thin tweezers, fine tip iron plus
solder braid) and the DP was no longer lit. Zero volts now shows
"00" instead of "0.0" and the display now
correctly reads our 3-digit HT voltage.
Is this chip based on something like a TLC5917ID?? |
 |
|
Switch-mode high voltage power supply
|
|
Next I'll discuss a cheap HT supply (suitable
for powering a valve radio) that uses a DC supply as its source
of input power. Because it converts DC to DC no particular problem
should be met by way of mains interference such as you'd find
in typical switch-mode AC to DC supplies which both radiate and
induce interference into the mains supply (where mains wiring
in turn radiates interference). It cost me £8.86 (=$11.27)
next day delivery post free via Amazon. |
|
 |
|
Above is the PSU in question. It's not too easy
to define the actual maker as there are no markings on the circuit
board, but as it uses identifiable parts (looking at the main
chip etc) I guess it's easy enough to duplicate so many people
may be turning these things out (note that this is missing the
fan connector and others one of the capacitors in order to maximise
profit). The switch-mode chip is a UC3842A and the circuitry
uses also an LM393 (a dual voltage comparator) a couple of VH01
(high power TVS diodes), an RU7088R (power FET) and a SOT chip
marked "8F" which I assume is a voltage regulator for
the input supply (quoted as 8-32V). The transformer is rated
at supplying 45-390V with an overall power rating of 40W and
input of up to 5A. I've printed the supplier's information below
but one must read (and understand their implications) before
using the PSU.
For example what's the maximum output current at the
max stated 390 volts and how does the latter vary with output
current. Of course one must also look at the heat dissipation,
particularly noticing the pair of pins marked "FAN",
the relatively small heatsink and the max efficiency of 88%.
I have a high power regulator module which I bought
for supplying the heater voltage for my T1154 and this has a
heatsink which is rather small for its job. This works fine but
again only once the ratings and limitations are understood ie.
it needed a much bigger heatsink than the puny one fitted. |
|
 |
 |
|
My plan is to carry out some basic tests (voltage/current/power/RF-noise)
then to fit the module into a diecast box to screen the receiver
from the PSU's 75KHz-based noise and, if all goes well, add a
3 digit voltmeter display.
40W output at 390V equates to 102mA. The RU7088 is
rated at over 60A (at 10V) and 62W dissipation at 100C with the
words "large heatsink"
Also Drain-Source no greater than 65V hence
the note in the spec concerning short-circuit protection (ie.
fit a series protection resistor)
A minimum of 12% losses will be 4.8W so a 12 volt
input will draw 44.8W/12V = 3.7A or a 24volt input will draw
a little under 2A
Initial results were as follows. I did notice a couple
of strange things. One was as the input voltage was increased
the input current seemed to get unstable although the output
voltage remained steady. Another time, as the output voltage
was raised slowly it suddenly rose to over 500 volts. Turning
off the input power then back on corrected this. During tests
the heatsink remained cool but the transformer got quite warm.
Best efficiency was achieved with higher input voltages.
One advantage of using this type of power supply is the steady
output voltage so one won't get the large downward swing associated
with a standard HT supply as valves warm up.
A receiver such as the R1475 whose power requirement
is 12VAC/270VDC can be run from a small 12 volt or 24 volt centre
tapped mains transformer. The HT negative from the module is
floating so can be used for the R1475 (or R1155) bias connection
without the problem of heater return connection to chassis. |
|
|
INPUT VOLTS |
CURRENT mA |
POWER WATTS |
LOAD OHMS |
OUTPUT VOLTS |
CURRENT mA |
POWER WATTS |
EFFICIENCY |
WASTE HEAT |
|
12 |
600 |
7.20 |
15,000 |
300 |
20 |
6.00 |
83% |
1.2W |
|
12 |
900 |
10.80 |
15,000 |
370 |
24.6 |
9.13 |
84% |
1.7W |
|
12 |
880 |
10.56 |
4,700 |
200 |
42 |
8.51 |
81% |
2.1W |
|
12 |
2,060 |
24.72 |
4,700 |
300.7 |
64 |
19.24 |
78% |
5.5W |
|
16 |
1,390 |
22.24 |
4,700 |
301.1 |
63.7 |
19.29 |
87% |
2.8W |
|
18 |
1,230 |
22,14 |
4,700 |
302.2 |
63.9 |
19.43 |
88% |
2.7W |
|
18 |
1,040 |
18.72 |
4,700 |
271.4 |
57.4 |
15.67 |
84% |
3.1W |
|
12 |
601 |
7.21 |
3000 |
130.9 |
42.7 |
5.59 |
78% |
1.6W |
|
12 |
805 |
9.66 |
3000 |
154 |
50.1 |
7.72 |
80% |
1.9W |
|
12 |
1,363 |
16.36 |
3000 |
200.5 |
65 |
13.03 |
80% |
3.3W |
|
16 |
1,400 |
22.40 |
3000 |
245.5 |
79.1 |
19.42 |
87% |
3.0W |
|
16 |
2,250 |
36.00 |
3000 |
299.3 |
95.9 |
28.70 |
80% |
7.3W |
|
18 |
1,997 |
35.95 |
3000 |
302.2 |
96.6 |
29.20 |
81% |
6.8W |
|
18 |
2,750 |
49.50 |
3000 |
353.6 |
112.6 |
39.82 |
80% |
9.7W |
|
|
Tests were curtailed at the 40W output point
because my 3K dummy load resistor was getting too hot for comfort.
It looks like you'd need a 24VDC 2.5A power supply
in order to get the best efficiency at 40W output and this need
not necessarily be stabilised as long as the off load output
wasn't much greater than 30 volts.. How about the noise level?
I connected my spectrum analyser HT probe to the end
of the load resistor with the PSU running at 18 volts input/250
volts output. The readings are not quantitative, just representative
and extend way above 30MHz. Left picture is PSU off and right,
PSU on. The worst peaks are around 50dB to 60dB above ambient
noise. Let's see how this can be fixed by screening and filtering. |
|
|
|
|
pending.... |
A Volt/Ammeter
|
|
|
 |
|
I'm not 100% sure what I bought because the
descriptions are a bit misleading. The voltmeter may be 0-100V
and I'm assuming the ammeter reads up to 10A because the circuit
board is marked "SJ-2DCVA 10A30V". it came with a pair
of plug-in leads (red-black-blue and a red-black).
The two-pin will be for powering the circuit board from the lower
Power Supply (rated at 4-30 volts) and the 3-pin for connection
to the (upper) power supply under test (up to 100 volts with
a maximum current of 10 amps). The picture shows red-red and
the description is catchall (below)...
As per usual, the main processing chip (20-pin) has
its identification code erased but it also has an LM258 chip
and a 3.3 volt "7133" regulator (see spec at the top
of this page) The pair of LED displays carry the code "LSH2031AUR"
and there's a pair of micro-pots for trimming the readings.
Unlike the voltmeter at the top of this page it wouldn't
be straightforward to apply say 270 volts via a potential divider
(in the lead marked "Thick Red Line") as rating of
the blue feed line is probably also 100 volts max.
Not bad value though at £3.68 inc delivery. |
3 Digit DC 0-100V 50A/100A Voltmeter Ammeter Gauge
Dual LED Digital Voltage Current Panel Meter Volt Tester Meter
Amp Detector(Red & Red-10A)
|
|
TINY SA
I'd redesigned my "high voltage" probe which
I se to protect my DSA815TG spectrum analyser so decided to check
its linearity. I used my oscilloscope with 50 ohm leads directly
connected to the scope inputs. The Tiny SA was set to -7dBm output. |
|
|
Tiny SA Frequency |
Tiny by DSA815TG |
Tiny by Oscilloscope |
HV Probe by Oscilloscope |
HV Probe by DSA815TG |
HV Probe gain |
|
20KHz |
-14dBm |
na |
na |
-27dBm |
-13dB |
|
50KHz |
-12dBm |
na |
na |
-25dBm |
-13dB |
|
100KHz |
-10dBm |
233mV (0dBm) |
1.18mV (-46dBm) |
-27dBm |
-17dB |
|
200KHz |
-9dBm |
154mV (-3dBm) |
1.4mV (-44dBm) |
-30dBm |
-21dB |
|
500KHz |
-7.9dBm |
95mV (-7dBm) |
1.8mV (-42dBm) |
-35dBm |
-17dB |
|
1MHz |
-7dBm |
78mV -9dBm) |
2.2mV (-40dBm) |
-36dBm |
-29dB |
|
5MHz |
-6.7dBm |
70mV (-10dBm) |
2.7mV (-38dBm) |
-37dBm |
-29dB |
|
10MHz |
-6.4dBm |
60mV (-11dBm) |
2.2mV (-4-dBm) |
-37dBm |
-29dB |
|
20MHz |
-6.4dBm |
50mV (-13dBm) |
1.6mV (-43dBm) |
-38dBm |
-31dB |
|
30MHz |
-6.8dBm |
na |
na |
-40dBm |
-33dB |
|
40MHz |
-7.5dBm |
na |
na |
-40dBm |
-33dB |
|
50MHz |
-7.7dBm |
na |
na |
-40dBm |
-32dB |
|
100MHz |
-7.8dBm |
na |
na |
-41dBm |
-33dB |
|
|
What you see is a mixture of variations. Assuming
that the Rigol is the most accurate equipment you can see that
the Tiny SA isn't too bad meeting its displayed power output.
The oscilloscope shows wildly varying voltages but partly because
I'm not using its probe, however the high voltage probe with
its 50 ohm output is more consistently measured by the scope
isn't too bad from 1MHz to 100MHz. |
|
Component Tester LCR-T7
Sold under countless names
|
|
I bought this thinking it would replace my three failed previous earlier
versions and because it includes its own rechargeable battery.
There were no instructions (except on the sellers page, copied
below) and I inadvertently plugged the leads into the wrong (badly
labelled) sockets leading to rather mystifying results such as
identifying a transistor when only two leads were used and back-to-back
diodes instead of a transistor.
Another feature proved to be a bit annoying. When
the leads are shorted together its a signal to complete an auto-check
so finding a really bad IGBT started a slow auto-test.
Compared with its previous 9-volt version this meter
doesn't give all the details, for example, of a capacitor.
Some advertisers do not include a set of leads but
in fact my green lead plug parted company with its lead after
ten minutes of use so maybe those sellers didn't supply leads
due to complaints?
I checked an in-circuit zener which came up with a
4-volt reading instead of ten volts, obviously because its battery
is only 4 volts and the firmware is limited. Its coloured leads
don't match the display either.
The pin numbering is awful because it doesn't always
match the colour of the front area. Mine as you can see is blue.
Number 3 on my device is green and (to me) invisible.
I'm OK with "1, 2 & 3" but what do "K,
A & A" stand for? Well, I found an on-line manual and
these latter sockets are for testing zener diodes, and scroll
down to see the spec.. |
 |
|
Description:
- 160x128 TFT display; - Multi function key
- Transistor test area; - Zener Diode test area
- IR receiver window; - Micro USB Charging Interface
- Charge indicator LED
Features:
TFT graphic display Multifunction Tester.
Transistor Tester; - Automatic detection of zener diode 0.01-30V
- Self test with automatic calibration
IR decoder; - Support Hitachi IR coding
- IR waveform display; - Infrared receiving instruction
Other
- Measure results using TFT graphic display (160x128)
- a key operation; - Automatic shut-off (Settable Timeout)
-built-in High-capacity rechargeable Li-Ion battery
- Detection of the voltage of the Li-ion-160x128 tft battery
- Transistor test area; - Zener diode test area
- IR receiver window; - Micro USB Charging Interface
- LED charge indicator; - Chinese and English support
Warning: Built-in Li-ion Battery, it is strictly prohibited the
tester immersed in water, or near a heat source!
Warning: For your personal safety, please strictly comply with
the use of Li-ion Battery specifications and precautions!
Operating Instructions
1.1 Key operational definitions; Multi-function key has two actions:
l Short press: Press the key and not less than 10 ms, release
key within 1.5 seconds
l Long press: Press the key more than 1.5 seconds; 1.2 Power
on
In the power off state, short press the multifunction key, the
tester is turned on and automatically measured.
l Power on & measurement interface; 1.3 Detect transistor
In the power off state or the test is completed, put the test
element into the transistor test area of test seat, and press
the locking handle, short press the multifunction key, the tester
automatically measure, graphical display of measurement results
when testing is complete.
Warning: Always be sure to DISCHARGE capacitors before connecting
them to the tester! The tester may be damaged before you have
switched it on!
Warning: We do not recommend using the tester to measure
the battery! The battery voltage must be less than 4.5V, otherwise
the tester may be damaged!
l Component placement; Test seat are divided into transistors
and zener diode test area, detailed in 1.1 Description.
Package Included:
1 set x LCR-T7 TFT Transistor Tester
|
 |
|
|
