Testing of the Batchelder - Grampian replica

[EpicenterBryan]

Hey guys,
I wanted to let you know that I have one of ROLANDJAYS "Batchelder" - Grampian replicas in my hot little hands and will be doing some testing on it this week.

IMG_4120.JPG

We ran into a snag right off the bat in that the hole for the stylus was just a tad too small and I couldn't insert a new Transco stylus. I sent John several other random styli I had as well so he could re-work the tolerances to accommodate all the slight variations.

IMG_4122.JPG

We also discussed some minor changes to the upper connection block to minimize chance of shorting out wires when the head was mounted on a metallic mount. It will be simple enough to move the connector block forward, but is fine for this test unit.

IMG_4121.JPG

I also have a feedback board Mark designed specifically for the Grampian head that will also be used. Here I installed it in a chassis since it has a 120vac transformer. Also moved the trim pot to the front panel and added connectors on the back panel.

IMG_4118.JPG

Just to be clear as I start posting data, I have turned down the offer of compensation for my time. I'm sure many of you would do the same and would be thrilled to be able to play with, and learn via hands on experience with something you don't own. Although I made it clear I will do my best not to blow it up and return it in the same condition - I also explained that I would not be held responsible for damage. This does not mean I will be doing a destruct test, but I will not be shy in trying to see if we can hit a standard signal level on disc.

So what do I hope to accomplish?

I will go through the adjustment procedures Mark has published for his version of the Grampian feedback controller. I will attempt to establish a sensitivity value using Flo's test disc as a baseline. I will publish a Non Feedback and Feedback version of the frequency response, using RIAA encoded test files Mark has provided that have the correct RIAA phase relationships.

Also, I will attempt to do a disc-to-disc copy of several mono riaa encoded songs using Wayne's preamp boards. The idea here is to eliminate the RIAA decode from disc, and the iRIAA encode to disc entirely when cutting a mono song that was originally cut in mono on disc. The idea here is to see how the Batchelder sounds with as close to the actual audio source as possible (minus a RIAA decode, and an RIAA encode).

So there you have it for tonight.
Comments or things you want tested?

Bryan

[Sillitoe]

Awesome, I've been itching to hear more about this project!!!

[Stevie342000]

Awesome, I've been itching to hear more about this project!!!

Me too, but one small possible fly in the ointment, did you ask permission from the BBC to use its design? All the published BBC papers on the web with the permission of the BBC state clearly that they are for information purposes only. If they came after you do not think that using a different name makes the design different and just because its an old design do not think that it is up for grabs, the BBC may not see it that way.

The cutter heads are a BBC design, they had nothing to do with Grampian, Grampian did not even make them.

[rsimms3]

Wouldn't the patent on any designs be expired at this point? I did a quick search and came up with either 17 or 20 years as the life of a patent. Different standards for the time period that this design was implemented and coming from the UK?

[Jccc]

O.m.g! (Drool all over keyboard)

Thats some mighty fine metal work

[Stevie342000]

Wouldn't the patent on any designs be expired at this point? I did a quick search and came up with either 17 or 20 years as the life of a patent. Different standards for the time period that this design was implemented and coming from the UK?

There may not have been a patent as it was designed by the BBC for in house use, they would own the intellectual copyright, which is something all together different.

It does quite clearly say on BBC information sites and on associated sites that any papers are for information purposes and not for commercial purposes.

Grampian was licensed by the BBC to produce the heads as were many others who produced other equipment. Just because the BBC does not have a R & D department does not mean that they may not take too kindly to misuse of their intellectual property, it was and still is funded by public subscription.

All I am saying is that at least you should check if these guys come after you with their big fancy lawyers you may be eaten alive. Grampian does not exist any more but they were the OEM, not the actual manufacturer that held the license. In turn Grampians name may have been licensed the BBC for badging purposes. Moy were the manufacturer and although they may be technically none existent they still do exist and are located at one of the British Film Studios. There license will probably have expired, I am neither a patent or person of expertise on intellectual patent.

Then again the BBC may not care one way or the other....who knows unless you check.

[Bahndahn]

Then again the BBC may not care one way or the other....who knows unless you check

And they won't even consider whether or not they care until someone in their law team finds out about it. Even then, the slightest difference or variation can over-rule a patent/copyright. Typically companies get legal about this if it threatens their 'profits', so considering these heads are no longer being produced there is ultra-low likelihood of this.

Such a nice head here! I'm curious to see the results!

[Bahndahn]

All the published BBC papers on the web with the permission of the BBC state clearly that they are for information purposes only.

As I imagine, there will be no way of proving that the head will be sold– particularly if he chooses to initiate all sales independently. Without official public documentation of the sales or intent to sale [by a registered business etc., prices published explicitly for the item] than all is clear and fine!

[tragwag]

excited to see where this goes!

[mratx]

I think you're okay with copying the mechanical workings, since patent life is 20 years in the US. The issue is going to be if you infringe on a trademark or copyright, so you can't use their company or product names or any other trademarked terms or logos that they may have. Although to be totally safe you should make the head look slightly different, as you can copyright certain visual design elements (although I doubt that actually occurred with a lathe cutter head.)

It's similar to the car replica business, you can create a mechanical duplicate of a car, but you can't use the name or maker's name without a license, and you can't make an exact visual duplicate of many cars because that gets into trademarked elements versus patented mechanical elements.

Mark

[mossboss]

At worst it is a close clone at best its a new design.
If even a thread or the length of a screw is different it cannot be seen as a copy
Copyright applies to written material, a registered design will only be offended if something is an exact copy, than passed on as an original, hardly, the case here
What with all the sx clones?
What a storm in a tea cup!
Never relised we had so many legal experts here, may as well add me. ha
Best

[EpicenterBryan]

Hey guys,
A quick update on testing. Much less progress than expected. The wife had to head out of town for a few days, and I'm stuck with tending to Monster puppy. He likes to chew on everything in sight including power and audio cords. I have help baby sitting him tomorrow and should have more time to debug... With that said:

Puppy Monster.jpg

I had two great conversations with MarkRob this weekend about the board he designed for doing feedback on a Grampian head. We did a bunch of debugging via phone related to unexpected responses from the feedback box. It turns out, the board I have is not one that Mark actually built but we think is based on his design. It has point to point wiring and Mark never made one like this. This is a board that John sent me and I have made 2 changes to it myself as well. I removed the PC mounted trim pot for feedback gain and installed a header / cable and front mounted multi turn pot. I also added a header to run an LED and resistor off the 12VDC supply.

This board varies from the schematics Mark posted in that the op-amp installed is an NE5532 and not the OPA2134 noted in the schematics. The installed Op-amp is fine for this application but not one Mark normally uses. Also, the + and - regulators are 12 volt versions and not the 15 volt units specified in the schematics. Again, they are fine for this application but not what he specified. So anyway, it's not a board Mark built but is based on what he published. So I do thank Mark for taking the time to go over things with me on this board!

I've decided to take a step backward and verify that the rest of this board is wired correctly especially since we (Mark and I) found there is an interaction when adjusting the feedback pot with the main fixed gain signal path which should not be happening. And the roll off frequency is not 2.4Khz as specified in the schematics. It's possible a capacitor value is incorrect and or a wiring error is going on...

IMG_4124.JPG

On a good side note, I added the silicon dampening fluid John provided, and added an o-ring on the top fill screw so the upper part wouldn't leak. John said to add 3 to 5 drops of fluid. I used a syringe with built in plastic needle to dispense the fluid. Then I left it sitting upright on several sheets of blank copy paper over night. Now - well after 24 hours has passed and there is no leakage of fluid. That's a great sign!

IMG_4130.JPG

IMG_4131.JPG

Hopefully I'll have more useful data in a few days and test cuts.

For now, rock on people!

Bryan

[studiorp]

Which type of silicon fluid do you use ? How many centistokes ?

[EpicenterBryan]

Hey guys,
Time for another quick update.

There has been quite a bit of progress made behind the scenes this last week. However, I'm not ready to post any specifics.

Some times the journey is more interesting than what is at the destination anyway.

Here is what's been going on.
Before I start, I need to thank Mark for his endless interest in the project. Mark and I have exchanged not only E-mails, but several debug sessions via phone.

I mentioned there were some anomalies that at first glance pointed to the feedback controller board. The board is based on Marks published design, but was not built with the PCB design Mark also published. Marks published board design has a proper ground plane, and all signal paths are shielded by the ground plane copper. His published design for the bottom ground plane is shown below:

Grampian FB Summing Amp Bot.JPG

Although the board John sent turns out to be built correctly, it is built with no ground plane and has point to point wiring on both sides of the board. That is an issue since the op-amp and several of the input resistors are located very near the AC transformer. While debugging, there was some HF noise which was eliminated when I placed two bypass caps right at the power and ground pins on the op-amp. They were installed already, but on the end of the board where the regulators are and not right at the power and ground pins. I also replaced the op-amp with the one Mark recommended on the schematic. There is still the issue of 9mv of 60hz on the output even with input and feedback shorted to ground. This may be acceptable for now but I'm not thrilled about it. I may either have to put some shielding in there as a band aid, or rebuild the op-amp section with a proper ground plane.

I know many of you are interested in knowing what the anomaly was that sparked checking out the box so thoroughly. Well, at first I just connected everything up to the box and went through the calibration procedure Mark published. The outcome didn't make sense when checking at different frequencies with feedback applied. Eventually, I skipped using the box and applied a noise source to the head and read the response from the feedback coils directly. The feedback coil output didn't make sense.

Mark and I decided to look into the inductance of both the drive coils and the feedback coils. The problem is right now I don't own an inductance meter (but one should arrive on Monday). So I found an old school procedure I never knew about that can be found here:

http://www.dos4ever.com/inductor/inductor.html

This procedure uses a function generator with a 50 ohm output impedance (and it must be 50 ohms), a frequency counter, and an oscope. I also added a true RMS fluke 8050A in the mix so I had a sanity check on the RMS voltage as well. The short version is that you set the frequency generator at 20Khz to start, at a known PP voltage, like 1V. Then you adjust the frequency only up or down so you have exactly 0.5Vpp. Then you read that frequency and run it thorough the equation. There is also a second equation where you can input the DC resistance of the coil as well. I did both.

The results were not what I expected for the drive inductance from published numbers, and although the feedback output numbers are not published, my calculations looked way too high. And Mark agrees.

So Mark suggested I test this method to measure the inductance of the coil from the driver many of us are messing with. The numbers are in the ball park so it's possible this old school impedance measurement scheme may actually be correct. But I have a real inductance meter on order that should be here early next week. I'll use that to do the actual measurement.

Since the Feedback output from the head was so different than Mark or I expected, I have a different interface that I want to try when it also arrives next week. I'm 99.9% sure that what I have measured relating to the frequency response output from the feedback coils is correct but...

Having clued you all into the journey, I want to say this:

Whether or not there is an issue with the usability of the feedback output function in existing form, it really doesn't directly relate to the functionality of this head in open loop mode. While we step back and look at things in a new light next week, I will press on with actual cut response test this weekend.

Having said that - John is sending me an original non-molested Grampian head to do some baseline tests with. I call this the "gold standard". I'll check this little puppy out so I have a baseline to compare how the replica compares. John is also sending a unit that has been disassembled so he could make measurements for his replica. There is a pole piece specifically that I want to take a look at in it's internals.

So there you have the update for now.
Not a very sexy update, but an update none the less.

Bryan

[dimi751]

Nice one Bryan looking forward to your testing
Does marks Grampian feedback controller have an
Inverse iraa Build in or is just for feedback purposes ??

[EpicenterBryan]

Does marks Grampian feedback controller have an
Inverse iraa Build in or is just for feedback purposes ??

No iraa, just for feedback.
-----------------------------------

So here is a super quick mini update...

John is sending the "Gold Standard" head and the model head tomorrow.

Most of today was blown messing around with configuring a swarf pickup tube since I'll try using a heated stylus for the first time on Johns cuts. I have used swarf tubes before but their position was never critical or really needed. In this situation it is needed to cool the heated stylus. So like usual I underestimated how long it would take to make something work. Thank God I have a 3D printer and could make a bracket to mount to my counter balance weight stuff for Groove Scribe.

IMG_4155.JPG

Love this bracket, and my 3D printer!

IMG_4156.JPG

You can't see in this shot but that is one curvy sexy swarf pipe if I have to say so myself. Damn nice.

That's all I have for tonight. Since I'm adding a heated stylus I need to make some mods to my vacuum pump wiring next, so the stylus power supply is only enabled when the vacuum pump is running. So many details to remember (or to forget).

Bryan

[EpicenterBryan]

Which type of silicon fluid do you use ? How many centistokes ?

That's a good question. I'll check with John. I think he said it was 100, but in reading some other info out there on the net I think I saw 200 mentioned as the recommended viscosity.

Stand by....

Bryan

[EpicenterBryan]

Here is another quick update:

I have to report to Jury Duty tomorrow. Since I'm the only conservative in a very liberal town, it shouldn't take much time away from new testing...

Let me first say this. The title of this post should probably be changed to "Testing of the Batchelder - Grampian inspired head". There are some physical differences between these heads and appearance. The physical size and mounting pattern is the same though. There are several improvements in my opinion externally that John has made, especially where wire connection is concerned.

On the left is an actual Grampian head, and on the right is John's version. You can see that really small screws are used in counter bored cavities for the wires to connect on the Grampian head and there are narrow channels for the wires and isolation to exit the back of the head. That really limits the size of the wire you can use. In John's version there are two sets of connections, the spacing for wires is larger so the risk of sorting things out is less. I have mentioned to him that moving the block forward is desirable to lessen the risk of wires at the back of the head from shorting to a metal head mounting bracket.

IMG_4175[1].JPG

On John's head, there is an additional row of cap screws. Those are connection points for the internal coil wires for both drive and feedback to get out of the head to the top for interconnection. The wires are extremely fragile. Yes, I have suggested that those should be soldered to the block so they could not be disturbed. The reason this came up is because in initial testing on the feedback output, the values were not consistent. Although John warned me not to mess with those screws because the wire was small and there was a risk of cutting the wire, I carefully tightened the connections any way, and problem solved. But how tight is too tight before the wire can be cut? I think that connection issue needs to be eliminated in the final version.

The reason that any chance of an intermittent connection should be eliminated is this:
Assume you apply 20db of feedback and have everything including power to your head set to your normal reference of zero ( what ever that is). Then all the sudden your feedback signal goes open. Your output to the head jumps +20db. Can the coils take that?

Here is a microscope shot that shows how small one of those wires is. As I recall, the head of that Allen cap screw is 3.5mm to give you an idea of size.

Still0005.jpg

As far as the internal guts go, there are two big differences. One is a major advantage, and one may or not be but is different. We will see how it goes when all tests are in. The magnets used in Johns head are the same size as the original ones but are neodymium and way stronger. The advantage is they never need to be recharged. Another difference in the formation of the pole pieces internally. Until I have time to do some tests on the original Grampian and compare the results, those two changes may or may not effect performance.

As you can see, John has set me up with a vast batch of items to compare. I now have an original Grampian to test, as well as one that I can take apart if needed. The second one has no charge on the magnets, and the feedback output is dead.

Then he also sent another one of his heads that can be taken apart if needed.

IMG_4174[1].JPG

As far as data goes (DC resistance, Inductance, Frequency response, Feedback output, Drive sensitivity and so on) for John's head... the first test in the can. I'm going to hold off posting it until I compare results with the original Grampian.

So that's the update for now. I hope to have a full weekend to take comparison data, then post results.

Bryan

[EpicenterBryan]

I'm at a point where I was ready to post data on John's head.
The final hurdle was to establish a drive level such that matches Flo's 1Khz test disc, and well as the CBS STR-100 disc. Then to cut music. I'll go into details about that in the next few days.

Oh, by the way... I want to thank Flo for sending me a copy of his test disc last winter. That is a really handy disc to have and is not expensive. It has the basics and can be counted on...

In the last 2 nights I made several cuts with reference 1Khz and pink noise to analyze. Earlier I calibrated a VU meter to read zero for 5cm/s lateral rms from the STR-100, but only for one channel of a 45/45 stereo cartridge - which is 3.54cm/s - track 6A/7A (one track is left only, one track is right only). I also did the same calibration for Flo's disc on another calibration path (jumper selectable) on Wayne's preamp board. I'm only using one channel - the left channel.

Here is how close my test cut came to the reference discs:

IMG_4183[1].JPG

Tonight, my most perfect level / noise test cut resulted in a slight loss in the HF response compared to an earlier incorrect level test from further outside on the 12" disc. By the time I got levels set ideally to match the test records, I was 1/2 way or more through the disc. A few operator errors like not recording the feedback signal for future processing didn't help matters.

IMG_4184[1].JPG

So this raises a question. I've seen "radius compensaters" mentioned before in presto catalogs and such. They have a series of contacts that switch in different resistors with capacitors and stuff to increase the HF boost as the cut gets closer to the center.

I don't have one. I bet most people don't either. So what do the pro guys do, and does VMS gear compensate for this? Perhaps Todd can comment?

As far as getting the best possible (most accurate) frequency response for John's head goes:

1) I could use the outer grooves of a 16" disc (which I have and can cut but but can not play back since my lathe has no tone arm). Does anyone have one they could loan for the project? Is that what was used when the original Grampian frequency response was established? If so, is it practical to use that methodology to evaluate John's head since no one would ever get that response at a smaller radius?

2) I could use the outer region of a 14" disc (which I also have) and I know I can play it back on my Pioneer turntable. I doubt anyone else can. Again, if the response is specked at that radius, how do we tell the user what he should expect with his stuff?

3) I could also use the outside of a 12" or a 10" (which I also have) and is probably more realistic for most potential users, but will clearly deviate from the pristine response quoted for an original Grampian head if that was derived at 16" diameter

I need some inputs from people about what is a reasonable diameter to test the final (published) frequency response at.

Help!

Bryan

[dubcutter89]

If I was selected to test this head I would do it like that:

Adjust setup to standard reference level
Cut a first test disk with discrete frequency bands
Don't use any radius compensation or something like that. No head eq! Feedback only!
Also don't do any eq to the signal.
RIAA (75us) off to check levels visually!

Beginning from the outside with:
1kHz reference, 15kHz, 12kHz, (maybe some more bands here to ckeck resonance),10kHz, 5kHz, 2.5kHz, 1kHz, 750Hz, 500Hz, 250Hz, 125Hz, 50Hz, 20Hz...
of course you can choose other frequencys or start higher

All testcuts on a 12" @ 45rpm, the more on the outside the better!

Maybe cut a second disc with 1kHz reference and some noise and a freq sweep, music, maybe also some different levels (+3dB,.., -6dB) for playback. Maybe also with riaa on. Watch current!

Keep in mind that a lot of measuring errors come from playback! for example scanning loss (hf loss at inner diameter, therefore starting with highs on outer diameter), or just the simple fact that nobody really knows the exact performance of his cartridge. Does it boost/attenuate highs? linear +/- 3dBs??
Immidiate analising/playback to avoid errors due to spring back of lacquer!

The first disk is light beam test and (ideally=good head) all different bands should produce reflections with the same width = same level
The second is for playback. Record straight into computer and compare the cut file with the recorded one and watch for differences.
FFT of noise will give a first look at what the response is like, but for a real professional measurement you would have to analyse the light bands with some special tools (or send the disk to flozki..).

Compare all heads.
I don't know an official freq. response of the old Grampians and I am very skeptical that they still hold that, or that 2 are same quality or whatever.
Also those old test were made with old equipment. Old blanks with different formula, old stylus, 78rpm?, ...nobody knows.

Best to show all results of all test, only this way the full information can be used.
Please don't post just a FFT of noise...

Hope this helps, all the best!
Lukas