Sunday, October 18, 2020

Ferranti 5820-99-795-7922 Noise Generator Coupler

 If anyone know the exact specification for this device I appreciate some comments.

Got this from the usual flee bay expecting to have a cheap noise source capable of 10Ghz since they are normally a bit on the expensive side.


Separating the two parts:

This is the power supply side:

And here the noise generator (underneath), a common noise source diagram:

Some similar Ferranti references, so this could be part of the same set:

Im assuming this is a 10 Ghz capable noise generator because of the waveguide and because Ferranti was involved in earlier radar set's at around the 10Ghz range.

End up doing an adapter/transition to SMA using just the noise module part, separating it from the waveguide:

Since I am a bit limited on 10Ghz test equipment the most I could do was connect it to an LNB and checking the level on an DVB-T USB dongle.

Bellow with power on at around 11Ghz (given the LNB offset of 9.750Mhz)

With the generator OFF: I know that there's some noise, anyhow could also be 1.7Ghz noise leaking on the SDR dongle.

I also tested using one LNB as amplifier for the noise (getting the signal before mixing) and then connected the output to the conversion LNB.

The results with the noise generator OFF (on the left) and ON (on the right):

Anyhow since I have no spectrum analyzer capable of 10Ghz, for now I assume it kind of works. I plan to do a filter around 10.3GHz and run noise trough it and then use the LNB and connect to an 1.5Ghz spectrum analyzer (or similar) for better conclusions.

Have a great day!

Sunday, September 20, 2020

DCF 77 Clock controller (hopf - HKW - clock mouse)

 I had a past test on this device but didn't touched it for around two years. Was convinced that it was a software or serial port situation. In the end nothing more than the antenna coil wire broken beneath the enamel (at the coil side) and one other that I had fixed previously at the PCB side. Because I tested continuity from the PCB to the antenna terminals only at that time I didn't realized the second situation until returning to it this last week.

Bellow the device:

And the antenna coil details:

After that sorted it works receiving the DCF 77 transmiter time. It's just a little picky in the direction and placement since I'm currently on the actual range limits. I used "vocap" site for  the actual direction of the signal.

There is a long thread on this device (in German) here. Also has a link for the "RCCD" software if you don't have original DOS software diskette provided with the device, it also runs under "dosemu" in Linux:

A native build for Linux can be found here (thanks to Guntram). I tested under Linux Mint 32 Bit. Bellow the debug output.

In the format HHMMSS and then day of week date year (better visible on the source code).

For 64 Bit Linux Mint you will probably get when running the hopf binary: "No such file or directory" when running "./hopf -d /dev/ttyS0" , you need to install the following lib32 "sudo apt-get install lib32stdc++6"

Inside the box:



The main chip for the RF reception is marked U2900, I could not find the datasheet for it but did found one that could be similar in terms of connections, as far as I reverse engineering it; the U2775B :



..still not sure if the same or not. 

Anyhow, I'm happy that the device is confirmed working and will go now to the box of the finished stuff without immediate use...!

Have a nice day!

Sunday, August 30, 2020

5LO38I CRT XY Oscilloscope

 Finished today the XY scope based on the 5LO38I CRT tube. It was working for some time but wanted to correct the voltage for the deflection amplifiers and missing a regulator, it was directly connected to the bridge rectifier, now has the required 7805 and 7905 regulators for the positive and negative voltages on this part of the circuit.

It's nothing special, it's based on previous experiments and adding now the deflection circuit. I have no space in the box to add a sweep circuit so it's kept as a generic XY display and not a full oscilloscope.

Because of the lack of shielding on the tube some re trace is seen, also does not help the lack of shielding on the vertical and horizontal amplifier input. As can be seen a bit here:

A bit more visible on the horizontal line.
Also during construction I had to reposition the power transformers because if near the CRT side will influence a lot on the trace quality. You notice that as soon as the power is taken, the dot became much sarper while the power supply caps still hold the charge.

Inside the unit:

The schematic is the one bellow except I added a 7805 and 7905 after the bridge rectifier to make the 5V for the op-amps:


Doing a Christmas tree with an Arduino:

... still had the inverted Y channel at this point.
Bellow the same on the Rigol DS1052E:

 That's it,

Have a nice day!

Sunday, August 23, 2020

GPIB - Rohde & Schwarz URY / URV5 Millivoltmeter

 Having the URY millivoltmeter decided to code a small utility to collect data via GPIB using Linux. As far as I know the URY is functionally the same as the URV5 being the URY a version for a specific marked, this software was tested only on the URY with Linux GPIB and Python.

The meter:

The probe used for testing:

The graphical output along with the launcher icon:

The code for the above software:

import Tkinter as tk
import pyvisa
import time
# read power from channel A of the R&S URY millivoltmeter
rm = pyvisa.ResourceManager()
inst = rm.open_resource('GPIB0::12::INSTR')

counter = 0
def counter_label(label):
  def count():
    global counter
    counter += 1
    result=inst.read_bytes(21, break_on_termchar='\r\n')
    result=result[9:] # remove the first 9 characters
    label.config(text=str(result)+" dBm")
    label.after(1000, count)

root = tk.Tk()
root.title("   R&S URY channel A  ")
label = tk.Label(root, fg="light green", bg ="dark green", font = "Helvetica 18 bold italic")
button = tk.Button(root, text=' EXIT ', width=35, command=root.destroy)

A screenshot in case formatting on blogger changes the code:

Made also a utility to get data via command line, could be handy for scripted collection. To get the values from channel B instead of A the code is only different on the instrument query string. The query is also only for dbm output, if you need any other unit or parameters do change the inst.write string.

Code for channel A:

copy/past version bellow:

# get the power of channel A on R&S URY milivoltmeter

import pyvisa
rm = pyvisa.ResourceManager()

inst = rm.open_resource('GPIB0::12::INSTR')
inst.read_termination = '\n'
inst.write_termination = '\n'
inst.timeout = 5000

  exception = 1
  result=inst.read_bytes(20, break_on_termchar='\r\n')
except: exception = 2

if exception == 2:


Output example:

The query string for channel B:

In case no probe inserted you will get the "URY PB NO PROBE" message

Usage example along with code and diff for channel B:

Hope it helps.

Have a nice day!



Sunday, August 16, 2020

Wavetek 3001 signal generator repair

 Got this Wavetek 3001 signal generator just for testing the FM deviation meter since I haven't here my other signal generator and it was the cheaper unit that could generate a signal changing the deviation I could find in the usual online flea market.

Was expecting it to be fully working but unfortunately the listing  didn't matched the reality. Instead of sending it back I decided to try an fix it.
It was not giving any output and the indication meter for the output signal was not moving, also touching the back right panel where the power supply dissipation is, it was getting alarmingly hot after some seconds of power on. Testing the power lines for the + and negative 18V it was showing close to 0V... so either the supply was dead or there was a short. Because it was the two supply lines I suspected initially the output transistors of the supply.

 Looking at the service manual and some web resources was easy to see the 18V power supply lines in short or not working could be due the usual short electrolytic capacitors.

Two great resources I could find with a lot more information than this post are the following ones: from K5JXH for the repair of a similar unit, a Wavetek 3006

and: by Ralph Llimek on a repair of also a similar unit, Wavetek 3000

In my repair the only hard part was to find which module is the one at fault, if not multiple. That was my approach because looking at the power supply I could not find any fault (top right part vertical PCB).



My approach was to removed 5 random ones and see. I choose the first 5 on the bottom from left to right. What I did also was to put labels on all the cables for future reference although after unbolting them and turning the box upside down they started to fall with the cabling attached, saving me some time. I was lucky the first time, with 5 modules removed it was back to life, that is, there was no longer a short on the power supply and the output metter moved, then I installed one after the other until find the culprit, it was module M29-2, see bellow.

The modules are easy to remove from the chassis, just unbolt the bottom screen and pull them from the top side.

Inside the module I could see a previous repair had been done on one of the supply lines capacitor because of the two different type, a more modern black on the right and the original one on the left. You can see also the small choke in series to the supply lines, positive and negative.

After measuring, it was confirmed as a plain short, so it was removed and replaced (not the same value but works anyhow):

 In the mean time I bough some more spares for future replacement.

I suspect it's also developing and issue on the 400Hz audio oscillator part, 1Khz is good but that will be for another day. Here's the two oscillators schematic in question:

Have a good day!

Sunday, June 21, 2020

AD584 and AD586 modules boxed for multimeter check/calibration

Nothing major here, just boxing some kit's/modules with AD584 and AD586 voltage references, enough for quick check of multimeters or any project needing a voltage reference.

The AD586 is a fixed 5V output wiht +/- 2.5mv of accuracy and the AD584 set to 10V with possibility of being change to 7.5, 5 and 2.5V with 5mv accuracy.

Connections are self explanatory, anyhow here's the AD586 with the metal lid covering the IC lifted:

And now the box calibrating the Range RE890G multimeter on the Volt range.
Itwas a guess work to find the correct trimmer since I could not get the tech manual for this multimeter:

Here assembled view from an older picture:
Checking another multimeter and my homemade milivoltmeter:

..more digits now:
...enough for most of the electronic work done around here.


Have a great day!

Sunday, June 07, 2020

GPIB - Marconi 6960A power meter

Another one for the GPIB series on instrument control from Linux, this one for the Marconi 6960A RF power meter.

The output on the screen (guess the power was lower than on the first picture):

Together with the milivoltmeter (and launch icons) getting data for recording, the milivoltmeter software/hardware is for another post and it's not GPIB, it's rs232.

My instrument is set on GPIB id 8 so you might need to change to reflect yours:
(inst = rm.open_resource('GPIB0::8::INSTR'))

=== code start
import Tkinter as tk
import pyvisa
import time

rm = pyvisa.ResourceManager()
inst = rm.open_resource('GPIB0::8::INSTR')

counter = 0
def counter_label(label):
  def count():
    global counter
    counter += 1
#   power=inst.query("SR3SR-TR10SQ1")
    power=power[3:] # renove the first 3 characters
    label.config(text=str(power)+" dBm")
    label.after(1000, count)

root = tk.Tk()
root.title("   RF power 6960A    ")
label = tk.Label(root, fg="light green", bg ="dark green", font = "Helvetica 18 bold italic")
button = tk.Button(root, text=' EXIT ', width=35, command=root.destroy)

=== code end

Have fun!