EnglishDutch Diagram BC-Tune D/A converter and amplifier
new BC-tune PCB layout - both sides
New ZIP file with example eprom file + NEW Visual basic 6 EPROM creation program - And a part of the program listing from VB6 - comments in DUTCH, with simple English Notepad HELP file.
Diagnostic output with over 1000 lines of calculation results.
Due to this list with tune info some minor roundup bugs removed.
If your XP or W2K is very 'clean' you also need
the VB6 RUNTIME and dialog files


Memory Modifications

Only a limited number of memory channels were originally used. From 1x 24 in the earlier NDH-515 to 4x 24 in the NDH-518. There are many more channels available, e.g. 256! Here some tricks to address them in binary mode and convert the display view back to decimal (in BCD mode). Eventually you get (max.) 8x 32 or 4x 64 channels!
If you change as less as possible on the outside of a NDH-518,
only one switch will double the available channels from 4x 24 to 8x 24.
Or for the NDH-515 with an extra (rotary) bankswitch from 1x24 to 8x 24!
In the inside mostly a chip for coding from BCD to binary and a second chip back from binary to BCD (for the correct display view) is always needed. This could be with a GAL array or an EPROM or both.
To address them all in an easy way also an additional UP/DOWN counter board was designed. But a more simple version with extra bank-switches is also possible.

The Up-Down PCB

Modifications executed on a NDH-515 memory unit:

This is the least complex version, only an 8 channel bankswitch is added and of course the binary encoding and binary-BCD decoding chip.
This will give 8x 24 channels in place of the original 1x 24.
If you look at the neat way it is added you wonder why JRC didn't do it this way.
PDF with binary coded decimal add-on universal (new) switch board for NDH-515
ZIP file with Gal array decoders - EQN and JED files
Circuit diagrams are for a slightly (minor) different GAL version as NDH-518
PDF with circuit diagram GAL expansion for NDH-515 memory unit


Sub page with
some pictures

UP/DOWN counter control for a NDH-518:

To address all channels in an easy way an additional UP/DOWN counter board was designed. The rotary 24 position switch has been replaced with an opto encoder. Like the tune dial!
PDF with up/down countercontrol add-on board PCB CAD
PDF with circuit diagram binary UP/DOWN counter control
PDF with binary to BCD decoding for displays with eprom - 64 channels p/bank
PDF with binary to BCD decoding for displays with GAL - 32 channels p/bank
Zip file with GAL related JED and EQN files for new 32 channel GAL version
64 and 32 decoding chips can also be used for the
(above) 24 channel rotary bankswitch version


Sub page with
some pictures

Accessibility modification on a NDH-518 unit to
make it into a remote controllable memory:


This is the MOST complex version, NOT advised to start with!!
PDF with circuit diagram part 1of 5 for memory accessibility options
(Memory control and read-out mod)
PDF with circuit diagram part 2 of 5 for memory accessibility options
(Up-Down control board with some additions for remote control)
PDF with circuit diagram part 3 of 5 for memory accessibility options
(Internal connections between sub-boards)
PDF with circuit diagram part 4 of 5 for memory accessibility options
(I/O buffer board and I/O connector wiring to external controlbox switches)
PDF with circuit diagram part 5 of 5 for memory accessibility options
(The channel control part of the external control box)
PDF with PCB CAD layout for memory accessibility options inside the NDH-518
PDF with PCB CAD layout control box for memory accessibility options
PDF with PCB CAD layout Up/down countercontrol add-on board
PDF with front panel example control box for memory accessibility options
Zip file with GAL related JED and EQN files for new 32 channel GAL version
(same GAL-chip and up/down CAD board version as NDH-518 above)

Sub page with
some pictures

Too bassy sound?

Early versions of the NRD-515 rev A and rev B had a bright, but somewhat distorted sound.
Revision C had the bassy sound, for most people a little too bassy.
It all has to do with the value of capacitor C333 (near TP-28), which is part of a low-pass network. In rev-B I've seen the value of 1 nF, which is wrong.
In rev-C I've seen a value of 33 nF ( = 0.033 uF) which is just a little too high.
The one and only correct value for the most perfect sound expirience = 22 nF = 0.022 uF.
If you change this capacitor C333, it is wise to change another cap C330, the one for side-tone IN. It does NOT change anything to the side-tone IN sound, but now you can (mis-) use this input for an external FM or ECSS-AM decoder.
I have used a value of 0.56 uF which was available.

AM detector buffer:

The single diode AM detector causes a too high a-symmetric load on the last IF stage.
With a small buffer with low Z-out the distortion is gone, also audible as cleaner SSB!!
PDF with drawing of an AM detector buffer
PDF with layout of an AM detector buffer



Alternative CW crystal filter:

JRC's narrow CW crystal filters are very hard to get nowadays and are too expensive.
Kenwood has equivalent 250 Hz - 455kHz filters, in the same footprint type YG-455CN-1, also from the NDK factory.
They won't work because the wrong impedance and a DC short circuit to ground.
But with two minor impedance transfer and DC separation toroids it will work,
even much better than the original JRC 300Hz CW-filter, which has a VERY bad shape factor (above 4) !!
See further down....


The toroids you see are 3E25 - 14mm - AL=2200
It transfers the 600 Ohm to 2000 Ohms and back.
It also gives the needed DC separation.
Windings are 20 for 600 side and 36 for 2000 side.
So it's about 1 to 1.826 in N, and 1 to 3.333 in Z.
A smaller 10 mm 3F3 toroid is even better. Less losses.
(recently available via TME-Poland)
Windings in this case 33 for 600 and 60 for 2000 Ohm side.



In a PDF with 3 sheets

CW FILTER COMPARE...
(A very bad shape)

The small curves left and right can be enlarged.
The RED curve on both is the 600 Hz KOKUSAI.
You see the curve shape of both filters
on two filterboard's.
The smaller GREEN curve left is the perfect KENWOOD
narrow CW filter YG-455CN-1, made by NDK.
-6dB =293Hz | -60dB =471Hz | shape =1.61
loss =4dB
At right you see the (GREEN) JRC CW filter YF455DPB curve. It is rather straddle legged and JRC should have
ordered a better filter at the NDK factory.
-6dB =329Hz | -60dB =1468Hz | shape 60/6 = 4.448
loss = 2.3dB
Shape 4.448 = very BAD ! Shame!

In a PDF with 3 sheets



Here is the link to the page how to measure those
filters and how to make additional buffer amplifiers
to connect it to a Selective Level Meter setup.
( Wandel und Goltermann !)
Here you see a quick preview of the setup.


Some available Filter tests in PDF
NRD-515 mainboard Rx-set1 - 6 kHz and 2.4 kHz
CLF-D6S and MF-455-10AZ121
NRD-515 filterboard Rx-set1 - 250 Hz and 600 Hz
KENWOOD YG-455CN-1 and KOKUSAI MF-455-03AZ121
NRD-515 mainboard Rx-set2 - 6 kHz and 2.4 kHz
CLF-D6S and MF-455-10AZ121
NRD-515 filterboard Rx-set2 - 300 Hz and 600 Hz
'JRC' YF455DPB and KOKUSAI MF-455-03AZ121
Some CFW455IT tests to make you dizzy (Dutch only)
Also DUO-tests with a selection of two to make 1 better filter.
NRD-515 filterboard Rx-set3 - 300 Hz and 600 Hz
'JRC' YF455DPB and KOKUSAI MF-455-03AZ121
Some quick test setup information - In English and Dutch now!!
NRD-515 filterboard2 Rx-set3 - 1.8 kHz and 600 Hz
'JRC' YF455EBF and KOKUSAI MF-455-03AZ121

HF circuit leakage around the filters:

After almost 100 measurements and the 4th build of the measuring module, the effects of HF leakage around a filter became very clear.
The effects are a.o. a too high bottom line and very straight, no curls or strong dips. Wait a minute, that's the same as the mainboard curves of the AM filter above!! So there is something not quite good in the circuit! After many tests it became clear that they should have used screened 330 uH and 4.7 mH coils. If you screen them now (with cut in half electrolytes as caps!) the L value and Q drop too much. They should have made it the same as the SSB filter, with screened transformers.
The difference is at least 20 to 30 dB!!
The small 14 mm 3E25 toroid Z-adapt solution I made for the Kenwood crystal filter is tested in the setup module and later in a NRD-515 before and after the 6 kHz AM filter, in place of 330uH, 220pF and the 4.7mH. So used two times. It gave 15 - 20 dB improvement. With an additional (still needed) screening bracket it even became 20 - 30 dB improvement!
Bad circuit is also used in other receivers, I've seen it used (in the diagrams) in the NRD-505, NRD-515, NRD-525, NRD-535, JST-145/245.Even Eska used it. I doubt if they used screened coils...

tested:
Max. -65dB is hardly possible with 'open' unscreened coils.
But with screened coils (o.a. with cut in half electrolytes!) and a fully screened setup I reached even under -100dB, so perfect is possible.
Tested with open 330uH Neosid coils (image): the aluminum cap reduces the L from 330uH to about 220uH, the Q drops from 45 to 16. So a coil with ferrite cap and additional screening is best.. Like the coils around the Kokusai filters.
I wonder how others build it... Screened? Or NOT?

The CLF-D6S original (brown) and with toroid modification
transformer 3E25 = green,
330uH 4A11 toroid = red
In a PDF
Transformer toroid:
14 mm 3E25
N-1K0 = 24
N-2K0 = 34


Test setup of the screened LC networks.
Very, very bad results without the caps!
Even with a 90 degree angle!
Image: Open coil test setup

Here (like above) how the Z-adapt toroids were used near the Kenwood filter.
Around the CLF-D6S it is still original.

View the NRD-515 mutilation with the transformer toroids.
Image: Transformer test setup

I'm not the first who found this design flaw with coil leakage around an IF filter. It has been mentioned earlier by Dallas Lankford. His proposal was to replace the 330 uH coils with coils on FT50-43 toroids. A toroid is well screened by design. But mind the 3x 90 degree non interfering position. (it can be improved still! I did with the transformer screening bracket! That's the thinner blue line in the above curves.)
Dallas Lankford's testing documents are hidden now on the net - not available anymore - secret?? Or do you have to pay?? I got a stolen version. Download it before they prosecute me.
I've also tested the toroid coils intensely, as well as in a test setup, as in the NRD-515 itself. See pictures.
Now I used a toroid with a much lower AL value, chosen is 4A11 - 13 mm (TN13/7.5/5), average AL = 360. Windings number is 31, 32 or 33 (measure!) With higher AL value the losses are higher and you have a greater L step by winding. Lower is better and more accurate.
Amidon - Fair-Rite FT50-43 were unavailable at the moment. Ferroxcube 4A11 is comparable, old color = pink.
During testing under 100 kHz (LC-resonance measurement setup with 10 nF in the ferrite-toroid setup module) the original 330 uH coils have at resonance at about 88 kHz a Q of about 80 - 85, the blue 330 uH Neosid testsetup coils about 45, the small 4A11 toroid is about 40 and the FT50-43 toroid is disappointing bad, about 6 to 7. The Fair-Rite core is therefore NOT adviced by me, it has much too high losses!
It is also possible that my deviant (bad FT50-43) toroids are from an earlier production date. I found this on the (former) site of Clifton Laboratories. One has the 'formula' of the composition of '43' sometimes adjusted (perhaps copied the composition 'formula' from Philips?)! See my ferrite toroid-test and measurement story over Here. It must therefore have been known by them that it was not perfect at all in the past.
Update: Meanwhile Clifton Laboratories removed this document from their site and is also taken over by DX Engineering ...
But HERE I have a copy on my server!!
The 2x 330uH coil in toroid shape gives easier a stable result compared with the transformer version above, 30 dB improvement is also possible. You don't have to ruin so much... It can again be improved by 5 dB more with an additional shielding bracket (the thinner blue line in the above curves), but is not needed.

Test setup of the 4A11 toroid networks.
In the copper foil: two cascaded Murata's.
Very good results without any coil cap!
Also here: at right (square) angles to each other.

Here the test setup of the filters on
the Rx board. Do you dare to do that???
I added additional screening!

Less NRD-515 mutilation with 2x the 330uH 4A11 toroids. This works great with minor destruction.
Mind the 90 degree angle!

Kokusai Filter Repair:
Kokusai Mechanical filters have a habit of deteriorating in time due to decomposing of the foam inside it.
It becomes a sticky gummy mess, which dampens the vibrations of the resonating disks.
With extreme care the shit can be dissolved with nail polish remover (in Dutch: aceton) or even better with isopropyl alcohol and repaired with fresh foam. After some experience: packing foam works better than cotton wool!!


Before the cleaning
Picture © by Luuk
Radiohobby

John Kolb ( KK6IL ) is a Kokusai Mechanical filter specialist and has a special site about it..
Go to John and read it over there in his own words.
John's site is NOT available all the time...
Due to public demand I've made a backup copy over here. Sorry in advance!

The pictures you see over here are from a radio friend: Luuk. He also managed to repair it too.
All the pictures © from Luuk in one small PDF sheet.


After the cleaning
Picture © by Luuk
Radiohobby


600 Hz - Before the cleaning
Picture © by PE1ABR

After a whole series of cleaning actions a friend of mine (PD0SDQ - Peter) developed an even better method of restoration. NO cotton wool, as I did like John, but packing foam.
Almost the same as on the pictures from Luuk. But square stiffer foam blocks with a hole in it over the piezo units, that is used at the ends to position the disks only, no pressure on the disks themselves. On top and underneath only a thin sheet of foam to fill-up.
Indeed it works just a bit better, lesser damping of the resonating disks!
Here you see what should be trick!!
One filter is a 2.4 kHz version, the other is a 600 Hz version. Do you notice the differences in disk numbers and the size of them??


2.4 kHz - After the cleaning
Picture © by PE1ABR

Frequency calibration:


PDF with frequency calibration notes for the NRD-515

PDF with example design (ref 4) of a frequency counter reference Droitwich PLL unit
PDF with design revision no 5 of the Droitwich PLL

Blown uPB426 (PROM) chip on the synth board ??
This chip is the band segment decoding table (PROM) out of the frequency range
Click to go to the page with PDF hardware information and
how to build your own decoding chip with a GAL22V10!!

VLF ADDON, extension down to 10kHz:
More modifications are needed to the input circuits to work properly
PDF with VLF mixer add-on (the extended version drawing)
PDF with VLF modification simple (overview sketch)
Excellent reception from about 25 kHz
Only exchange some capacitors and two resistor like inductors
PDF with VLF modification extended (overview sketch)
Only do this if you are experienced, also a relay is added and
an extra input toroid in the first mixer to extend the Rx range
to as low as 10 kHz without degradation of sensitivity

PHOTOGRAPHS section:

Photographs from the NDH-518 memory unit
with eprom BIN-BCD address decoder,
push-button and rotary-opto up-down control
Click here or on picture for sub-page
Photographs from the NDH-515 memory unit
with two GAL's for BIN-BCD decoding.
Only one hole for extra bankswitch needed
Click here or on picture for sub-page
Photographs from the NDH-518 memory unit
with two GAL's and remote control options.
Most complex version
Click here or on picture for sub-page
The design of a set of replacement uPD5101
modules for an exploded memorybank

NEW
In DUTCH and ENGLISH available
Click here or on picture for sub-page
Photographs from the NWG-515 Remote VFO
A NWG-515 ?!?, a very peculiar thing.
Click here or on picture for sub-page
Mains-inlet changes
Safety modifications mains filter to comply with new regulations
IEC power input exchange with real mains-inlet filter.
Click here or on picture for sub-page

Click over HERE for a direct hop over to all the other radio information on the next page.

Service Manual Bmw


If you do not have the 'Acrobat Reader',
please access to the site of Adobe Systems Incorporated.
The Acrobat Reader is distributed by Adobe Systems Incorporated free of charge.
Adobe, Acrobat Reader, and their logo marks are registered trademarks of Adobe Systems Incorporated. Gm serial numbers engine.

A nice NRD-515 YouTube video I found,
a perfect place over here to embed it.
And there is a lot more to see..

Copyright© 2001 - 2013 Walter Geeraert - All Rights Reserved

index515.htm
by Walter - PE1ABR - 2020-06-29