|
|
Last update: 03/06/01
13.15.00 |
Hallicrafter SX-28
restoration log
17-4-99
This is the kick off day. I've taken the pictures in the cellar using my webcam, and brought the radio upstairs in my lab. Taken some pictures again just to remember as it was…
Got the rig out of the box,
A first cleanup with vacuum cleaner, a brush and a dry damp cloth and compressed air can.
A lot of dirt but signs are all readable. S-meter needle is positioned at the end-of-scale on the right: it seems not normal. All commands seem to mechanically work well but the "crystal phasing" whose shaft is definitively too loose. Turning the tuning knobs the dial moves accordingly, but the left one some time is loose. The AF Gain rotates indefinitively and "sounds" suspect.
All knobs are in place and look original. Removed all knobs (no problems). Knobs are now in a bag with a note about their original position.
The main transformer has a big beautiful "h" on
it. It is the Hallicrafter logo: it is original!
The RF section is open: the removable plate is missing but fortunately all the
thumbscrews to fix it are in place. Tuning appliance partially working but a
metal wire is floating around. The chassis is brown in colour and covered by a
film of brown crud. It passed the magneto-test so it is a steel chassis.
Crystal phasing shaft broken. All tubes were in place. Tubes removed and
cleaned with the dry cloth only. Entries created in the tubes sheet of
the RDD. All tubes have been put in separates plastic bag with extensive
description written on them (ID, what it should be, what it is, socket it was
removed from,...).. Some of them have signs not corresponding with Halli's
schematics. A strong compatibility and functionality check is needed.
BFO's shaft is really loose: it seems to be due to a mechanical problem of the
variable cap ..
Visual analysis downside (Chassis image1, image2 - 1,2MB each!)
Some of the caps were already replaced. Wiring seem
partially damaged and in some point not original. Bandswitch appears in good conditions
even if some contacts seem heavily oxidised. All tuning trims in sight look as
moved - the red fixing painting on them is not intact. Many have the coil's
screw cut broken, one has a little bolt soldered on top - this does not look to
be an original arrangement.. Found cold Solderings of coil springs:
springs move with coil rotation. The rig has already been aligned by somebody
else (Who? Where? Why? How?).
Selectivity switch is almost wired with "modern" wires (plastic):
evidence of tampering with.
Electrolytics not original; Magneti Marelli 2x(16+16uF 525Vp). They look good.
Go to the top of this page | Back to the Sx-28 page
24/4/1999
After a lot of readings on the net and on some books, I am back to the radio.
Started working on power supply. Removed CH1, electrolytics and T8.Wiring check; SURPRISE:
R43 e C76 are missing; there is an empty lead
in z20a that is probably there to support their wiring.
C47 missing; Found R42
modified by parallel of two 450 ohms and wired in very "strange" way;
CH1 successfully passed continuity tests and ohms test
T8 successfully passed continuity tests but has spoilt leads. Ohm test NON PASSED: measured value are far away from specs. By sure this transformer is not original. Some action is needed.
T7 (power) windings successfully passed the continuity test. Warm test passed (bulb VARIAC used) ; all voltages within limits; insulations OK.
Go to the top of this page | Back to the Sx-28 page
14/5/1999
Started working in the power supply and LF amplifier areas. Intensive components identification and wiring understanding. All components have been identified and recorded on the RDD (with some picture of the original wiring).
Go to the top of this page | Back to the Sx-28 page
22/5/1999
I have been to a local hamfest and I've bought a new digital multimeter also able to measure capacity
Go to the top of this page | Back to the Sx-28 page
29/5/1999
Started working on the power supply and the LF amplifier. Involved tubes are v15 (rect.),v14,v13(final PP) and v12(pre&inverter).
Resistor are almost all out of value: variances grater than +20%; some more than 50% . Tubular paper caps look very bad. To remove some components (resistors and tubular caps) I had to remove others (mica caps) that are not necessarily bad….
The decision is taken to rebuild the LF amplifier from
scratch.
All components identified and taken some additional picture.
This is a snapshot of my component
database. Note the "remove sequence number":
it represents the order by which I’ve removed the components – to be used as a
guideline when reinstalling new components. Wiring check against weak
insulation. These
are the LF-amp
and power
supply diagrams I used working at restore time.
Intensive cleaning of tubes sockets by mean of a teeth
brush, water and a little of alcohol (not so effective..). I also removed and
cleaned those "lead arrays" (I
don’t know the correct English term for them) so common in vintage electronics.
Go to the top of this page | Back to the Sx-28 page
10/6/1999
A decision has been taken not to maintain the original appearance for LF-amplifier (new components will be in sight under the chassis - horror for somebody, I know!).
A couple of non standard value caps recreated with pairs series. Non standard resistors values like 200K and 250K got by measuring actual value of some NOS resistors I have in my sliders since a long time (e.g. an old 220Kohms resistor low in value is used to replace a 200Kohm resistor).
Each component back to its original position. The sequence
numbers and the pictures I 've taken before dismounting saved me a lot of time.
Next step will be to reconnect the power supply, disconnect B+ from the other
circuits, fire up the LF amplifier and test it.
Go to the top of this page | Back to the Sx-28 page
July 1999
I have sent all the tubes to my old friend Mario (IK1HXN) for testing. He has a friend owning a tube tester.
Looked for and found a company specialized in rewinding transformers. I have given them the T8 specs and they have provided me a new output transformer for 20$.
September 1999
The tubes came back to me with the test results: they are almost all bad. This is the tube sheet with the tests results:
|
ID |
required model |
Function |
found inside |
is it pin2pin compatible? |
Status Sep 99 |
|
V1 |
6AB7 (1853) |
1st RF Amplifier |
6AB7 (1853) |
Y |
OK |
|
V2 |
6SK7 |
2nd RF
Amplifier |
6AC7 |
Y With better pendenza |
OK But Very low |
|
V3 |
6SA7 |
Mixer |
VT-150 (=6SA7) |
Y |
OK minimum |
|
V4 |
6SA7 |
HF Oscillator |
6SA7 |
Y |
OK |
|
V5 |
6L7 |
1st IF Amplifier Noise Limiter |
VT-87 (=6L7) |
Y |
OK minimum |
|
V6 |
6SK7 |
2nd IF Amplifier |
6AC7 |
Y With better pendenza |
OK But Very low |
|
V7 |
6B8 |
2nd Detector S-meter driver |
VT-93 (=6B8) |
Y |
Inefficient |
|
V8 |
6B8 |
AVC Amplifier |
6B8 |
Y |
Inefficient |
|
V9 |
6AB7 (1853) |
Noise Amplifier |
6SK7 GT |
Y But different xx pendenza xxx Try to use |
OK Minimum |
|
V10 |
6H6 |
Noise rectifier |
6H6 |
Y |
OK But Very low |
|
V11 |
6J5 |
Beat Frequency Oscillator |
6SN7 GT |
N |
OK minimum |
|
V12 |
6SC7 |
1st Audio
Amplifier |
6SC7 |
Y |
OK But Very low |
|
V13 |
6V6 GT |
Push-pull Audio Amplifier |
6V6 GT |
Y |
OK |
|
V14 |
6V6 GT |
Push-pull Audio Amplifier |
6V6 GT |
Y |
OK minimum |
|
V15 |
5Z3 |
rectifier |
no writing |
? |
OK |
|
?? |
6H6 |
Found inside the rig |
No writing |
Y |
OK minimum |
So I have put an order to ESCO elettronica (TODI -PG Italy) for the following:
2x 6B8, 1x 6J5, 1x 6SC7, 2x 6V6GT
Go to the top of this page | Back to the Sx-28 page
November 1999
Received the ordered tubes and , surprise!!! 6B8, 6J5, 6SC7 are glass, while the ones that were in the rig were all metal.
Raised a question to the BA community (look for "glass Vs. metal tubes" on rec.radio.boatanchors)
December 1999
Bought a Tektronix 475 oscilloscope and a National function generator VP-7402A.
Worked a lot on building up this site.
Go to the top of this page | Back to the Sx-28 page
15-01-2000
Restored the ST-2 and ST-3 speaker connections. One of the insulator was missing and has been reconstructed starting from the plastic battery cover of a broken toy (it was a cellular phone dummy). These are all the pieces just before remounting them. The two connectors are now in parallel with the same output impedance=8ohm. Restore the phone jack also (some cleaning and functioning check).
Go to the top of this page | Back to the Sx-28 page
20-01-2000
The entire power supply has been rewired; filter caps mounted on the original clips. GND connection via a single wire to the chassis (wiring not original-see discussion "help on some SX-28 wiring" on rec.radio.boatanchors.
Go to the top of this page | Back to the Sx-28 page
14-02-2000
Everything is ready to fire up for the first time. Done using my bulb variac. Once reached full throttle the LF amplifier is still silent.
Measured some voltage around … seem normal, but the grid of the first section of the 6SC7 tube seems floating. I now remember that R33 (500Kohm/log variable resistor) , the AF Gain regulator, was tricky during the first glance I had to the rig. Measured with my VOM: the central lead resulted open. I have open it for inspection and found it almost destroyed inside. Found a replacement.
Turned-on again…. IT WORKS!! The loudspeaker is emitting the characteristic hum when touching the input line from the PHONO jack with my fingers, very low white noise and no hum otherwise.
Put the output of my Kenwood TS-430 into the phono jack, and I was able to ear quite a good sound; but…..
The 6V6GT were getting hotter and hotter, until reached a temperature that sounded not normal to me. Better investigating with a signal generator and the oscilloscope, I discovered a very low output level and high distortion (upper wave not equal to the lower one..). The output signal levels of the 6SC7 were in good shape, in opposite phase but of different levels (30Vpp of difference).
Go to the top of this page | Back to the Sx-28 page
20-02-2000
During further checks I understand that all the NOS out of value resistors I have mounted on 10/6/99 were gone completely out of value - even 60-70%. That day it was a very bad idea to use NOS resistors! I have 30 of them so I made a test: I have measured the value of one of them as taken out of my sliders; it measured 227Kohms. Than I simulated soldering it warming it up with my iron. When checked it again it measured 378Kohm, even after getting cool!!!! I junked all of them and I had some hard time in replacing them in the rig with new stuff. Here is how they look like - if you see any of them stay apart!!
After the work was done, almost nothing changed in the amplifier behaviour - the 6V6 are really to hot, and I cannot figure out why. I getting more and more frustrated…
Go to the top of this page | Back to the Sx-28 page
6-3-2000
After days of checks, I finally discovered one mistake: I have mounted a 22ohms resistor for R42 instead of 220 ohms. To be in a hurry is always a bad approach! With a 22 ohms bias resistors the bias for the 6V6s was as low as 2Volts instead of the 15V I expected to have. I hope not to have ruined my poor new 6V6s…
Now that the problem is fixed the power amplifier sounds very good and I cannot see any distortion anymore using my test equipment.
I'll never get confused again by the black and brown stripes on resistors - honest!
Here are the measured
voltages within my unit:
|
|
test 0 |
Test1 |
Test2a |
Test2b |
Test3 |
|
Mounted
tubes: |
none |
V15 |
V15,V12 |
V15,V12 |
V15,V12,V13,V14 |
|
measurement
condition |
T7 outputs |
|
sw10=out |
sw10=in |
sw10=in |
|
T7
primary l5-l9 |
108 |
107 |
107 |
107 |
108 |
|
T7 secondary
l3-l11 |
6,4 |
6,3 |
6,3 |
6,3 |
6,2 |
|
T7 secondary
l6-l10 |
5 |
4,7 |
4,7 |
4,7 |
4,7 |
|
T7 secondary
l24-l12 |
600 |
596 |
594 |
594 |
590 |
|
T7 secondary
l24-l8 |
300 |
299 |
296 |
296 |
294 |
|
T7 secondary
l8-l12 |
300 |
299 |
296 |
296 |
295 |
|
C49 |
|
357 |
360 |
360 |
264 |
|
C48 |
|
355 |
351 |
351 |
255 |
|
R36/R37/R38
common point |
|
348 |
242 |
242 |
177 |
|
V12 p3
first grid |
|
|
|
|
0 |
|
V12 p4
(inverter grid) |
|
|
|
|
0 |
|
v12 P2
(first plate) |
|
|
249 |
172 |
178 |
|
V12 P5
(inverter plate) |
|
|
189 |
189 |
130 |
|
V12 P6
(cathode - bias) |
|
|
2,13 |
2,13 |
1,6 |
|
V13 p5
(grid) |
|
|
|
|
0 |
|
V13 p4
screen |
|
|
|
|
255 |
|
V13 p3
(plate) |
|
|
|
|
252 |
|
V13 plate
current (mA) |
|
|
|
|
33,6 |
|
V14 p5
(grid) |
|
|
|
|
0 |
|
V14 p4
screen |
|
|
|
|
255 |
|
V14 p3
(plate) |
|
|
|
|
253 |
|
V14 plate
current (mA) |
|
|
|
|
31,5 |
|
V13&V14
p8 cathodes |
|
|
|
|
14,7 |
Now I have to put CH1&2 in their original position - then the Power supply & Lf amplifier stage can be declared OK-
This site has been transferred for the first time to the ISP (Telecom Italia Network - free service) server for publishing. It was very hard work, but now I am on-line. Due to performance and service problem I have not yet decided if I'll stay with this provider in the future.
Go to the top of this page | Back to the Sx-28 page
14-7-2000
Started working on the BFO control variable capacitor (C57). It is in very bad conditions mainly for excessive wear and tear. The bush where the shaft rotates in, is so worn-out that the shaft has so much side clearance to get the plates in touch. The shaft has also a significant end float. Definitively the solution is to replace it, but willing to exhibit some mechanical ability, I tried to fix it and here is the story.
First I have dismounted it into its primitive components. The lack of mechanical strength between the main shaft and the shaft cover made the job very easy, but it also was the main reason for the shaft clearances. I noticed that, being the shaft forced inside, the bush still allowed the shaft to stay in a correct position, keeping plates apart for the full rotation. So I focused on restoring the mechanical strength necessary to maintain the shat in the correct position. Originally this strength is provided by the shaft cover. The idea was to replace the lost pressure force, threading the shaft and install a nut and counter-nut play . Good job.
Now the problem is how to reinstall the shaft cover, now that the shaft diameter has been significantly reduced as a result of threading… OK soldering by mean of a handy pencil torch. Success!
The net result is a working variable capacitor, almost original, but with one problem: the rotation is not as smooth as it should be. The reason has to be looked for in the friction between the bush and the nut. In fact both the bush and the shaft cover (the components that were rubbing in the original set up) are made of brass, while the nut is made of steel. Lesson learned: steel and brass do not rub very well. I don’t expect this solution to be the definitive one, but it gives me time to look for a good new replacement.
These are all the pics of
the C72 restructuring
|
|
|
|
|
|
|
|
|
|
|
|
Go to the top of this page | Back to the Sx-28 page
3-10-2000
Started working on the BFO circuit. The tube found inside my unit was a 6SN7. It should have been a 6J5. A friend suggested somebody could have used one section of the 6SN7 as a replacement of the 6J5 triode. BINGO!! The wiring was a real mess, absolutely not in line with the Halli diagrams. So I decided to remove all the components attached to V11 and T4 and to rebuild the BFO from scratch. Just for the sake of history, here is the removal log:
- not original Geloso
electrolytic Cap (10uF, 30V) connected to: (-)T4-GND lead, (+)V11pin3
- 15Kohms resistor between
V11pin8 and V11pin3 (reads 16K)
- original SPRAGUE paper cap of 0.01
MFD 600V (ex C74) between (external foil) T4GND and V11pin5
- wire running from the
bandswitch wafer nearby to V11pin8 (unsoldered only from the tube socket
side)
- wire running from B4pin2(GND)
to V11pin8 (the previous wire was only used to bring GND to the switch –
needs double check later)
- 20K 10% 2W carbon (ex R45) from V11pin5 and B4pin1
- 50Kohm resistor (ex R46) from
V11pin8(GND) to V11pin4
- 250Kohm from V11pin3 and B4
pin2(GND)
By the way, by comparing the BFO diagram with the 6J5 and 6SN7 pin diagram, one can understand what some of the crazy previous owner attempted to do.
Here is the empty V11 socket ready for rebuild according to the original Hallicrafters design.
Wanted to check T4 conditions. Dismounted T4 loosening the two nuts between V11 and V12 sockets. Extracted the coil out of the shield by loosening the top nut. It looked like this; please note that the little threaded tuning shaft was damaged: one half of the top cut is missing. First of all I dripped some new wax (leaked from a candle) to fasten the coil and soldered a new top to the tuning shaft. I used a kind of brass plug (pic1, pic2, pic3) often used in house furniture’s (like bookshelves) to create threaded females (hole) to screw in metric screws in the wood.
Before starting rebuilding the BFO, I started working on the detector/S-meter/IF-AVC stage. The driver for this choice being that V11 and V7 are so close each other that is seems to be convenient simultaneously working on both.
Here is the components removal log:
- identified the three leads of
the screen grid divider: pin1&pin2 being the R32 leads and pin2 and pin3
the R31 leads (in fact pin1 is connected to B+ line and pin 3 to GND – as
per diagram)
- to be able to work into the V10
(6H6) socket,
- dismounted R29, the s-meter’s
zero setting potentiometer;
- disconnected R28 (a 100 ohms
resistor) from one external lead of R29 to V10pin6 (unused in 6H6);
- disconnected the non original
resistors coming out of the other external lead of R29 and directed to
the screen grid divider pin 1
- disconnected one side of a
0.03uF 600V paper cap from V10pin1(GND) ; the other side is at T5
- to get better access to V7
dismounted the ANT. TRIMMER shaft by unscrewing the two screws in the shaft joint close to C6
still to get better access, I also decided to get rid of the bandswitch shaft…. But just withdrawing it without pulling it back enough to get out from even the first wafer; to do this:
· loosened the screw on the brass cylinder installed on the shaft immediately behind the front counter panel
· unscrewed both screws fastening the shaft to the back of the chassis…
· then very carefully hit the shaft with a lite hammer, forcing it to move backward; SIMULTANEOUSLY addressing its back gear out from the hole in the back of the chassis. I’ve been very careful not to withdraw the shaft too much: the head has not to leave the first wafer: I don’t know, at the moment, what can happen in the wafers in case the shaft is removed out from them.
I was afraid about the two little balls contained in the shaft gear and about loosing synchronization between the shaft rotation and the shaft detents. Fortunately the shaft gear remains intact and you don’t occur in such a risks.
Go to the top of this page | Back to the Sx-28 page
15-11-2000
Had a strong look at the
ARRL Handbook about the AM detection process, AVC development and noise
limiting. Redesigned the detector/ANL/AVC diagram as it can be found within my detector page for better understanding.
10-12-2000
After days of hard work, I have removed all the components and logged my activity in the component sheet. Note the removal number: it represents the order by which I’ve removed the components – to be used as a guideline when reinstalling new components. Also removed the V11 socket. The V7 socket is still in chassis as it has a lot of wires very difficult to get rid of. I’ll see.
I have also created the chassis diagram ; it is very useful! Hope this will help other people to save time in diving within the sx-28’s chassis. It would be a good idea to do the same job for the power supply and the LF amplifier… I’ll see..
Go to the top of this page | Back to the Sx-28 page
25,26/12/2000 (Christmas’s
days)
Finally removed V7 socket as well as V11. Cleaned up wires. Wires repainted using PAPER-MATE permanent markers. Everything is ready to start reconstruction.
Dismounted T3. As it looked very “dry” with some flaws in the protective material and trim caps oxidized and dirty. I filled it with some silicon to protect and secure the coils. Trim cap polished with G-20 to deoxid and clean them up.
To start rebuilding the BFO I designed a possible layout for the components as, due to the heavy tampering with suffered by my unit, I don’t have a reference for it. Posted a request of help on the [email protected] mailing list.
Due to this waiting status, I decided to play a little with T3 and my signal generator and scope. Connecting it like this, feeding the primary coil with a signal sweeping between about 400KHz and 500KHz. I have got this nice behavior (values are indicative only as - incredible to say – I miss a frequency counter! I only used my Kenwood ts-430s as a raw selective voltmeter tuned to 455KHz. Furthermore my sig-gen is not very suited to weeping between frequencies: it only allows for sweeps in a range expressed by a ratio of the main frequency…). Modifying the connection like sketched in the bottom part of the diagram, that is adding a “detector” to the circuit, I have got this nicer picture as expected (the two side peaks are just the noise produced by sig-gen at the end of a sweep sequence..). Trimming on the primary cap the curve improved a little in its shape getting narrower. Trimming the secondary the effect was to move the peak around the frequency range. By the way I have got the same results either with direct sync between the sig-gen and the scope working in X-Y mode, or with the scope in normal mode and adjusting it to stay in sync with the sweeps.
Definitively a good time for my Christmas vacations!
Go to the top of this page | Back to the Sx-28 page
10-1-2001
Rebuilt the BFO circuit and now ready to fire my rig up again (it’s a long time since it received power…).
Surprise: some arching happened within one of the 6V6… so I have immediately shut down power, pulled out all the tubes and started investigating the reason for this strange behavior. Giving power again with all the tubes out but the rectifier(5Z3) I measured an unreasonable high potential in the B+ line (360V!). My memory served me well in remembering that back in march, at the time I gave it power to test the LF amp, I had some difficulty in getting stable volt measurements.
My suspects go toward the R31-R32 tapped resistor.
Go to the top of this page | Back to the Sx-28 page
|
|
Last update: 03/06/01
13.15.00 |