A good reason not to trust color codes.

[Jerry]

I finished up my second TGSL a few days ago. This one is the EDU version. It worked right away when I powered it up, but was lacking in depth and something was not right with the response.

After some testing I found that the filters were centered at about 5 hz and things dropped off fast above 10 hz. Also it took a good millivolt injected at the source of TR4 and TR5 to drive the LM308's to about 350 mv.

I went back to my original TGSL and found the filter response and gain to be much different so I started checking for proper component values. I had done this before, or so I thought. Anyway, I found that I had installed 47k instead of 4.7 k resistors for R41 and R53.

After installing the proper parts, I repeated the same test and found the filter was now centered on 10 hz and the gain was very hot. With the signal generator turned to show zero volts I was not able to see the input signal on one channel of my scope with the gain wide open. However the LM308's were still showing the signal until I turned the attenuator all the way to off. Impressive sensitivity!

Here is a picture of the offending parts. I included a 4.7k resistor for comparison. It is the one on top. My eyes have trouble with the multiplier band. So the only way to know for sure is to measure.

Jerry

P.S. When I sort everything out with my test of the filters, I will write it up and post it somewhere. I think it will be handy to get some baseline information so people can use it for troubleshooting their TGSL projects.

[Davor]

You have my sympathies. Recently I switched to metal films and they all have an extra band and it is not gold so extra care must be taken. My solution is buying them in batches and sticking them on pages in a folder. I always have all the values from E6, at least a few each, and I tend to use E6 values in my projects. I check them with multimeter before they get there. Some values need replenishing more often than the others. It may look a pricey solution for a start, but pays off very soon. You just browse through your folder and it is there.

[Jerry]

I have been buying them in batches of 100 lately which helps because they are all on a continuous tape. However I still have large quantities that are loose and in plastic bins.
I like your idea but for me it is a matter of getting it organized.

Jerry

[satdaveuk]

I have been buying them in batches of 100 lately which helps because they are all on a continuous tape. However I still have large quantities that are loose and in plastic bins.
I like your idea but for me it is a matter of getting it organized.

Jerry

Hi Jerry
Got cought out on that one with high stab 5.1k 2 watt resistors years ago from thorn which was part of the BRC in them days on a TV repair, since then always check each component before installing, which falls into good practise now, because we are not looking just for accuracy of values of components but also tolerence.
As ive said in previous posts out of every batch of 50 caps, your lucky to find 5 that fit the bill if your goal is high priscision, which in my eyes is needed in at least 45% if not more of each circuit in metal detector builds.
Metal film resistors are pretty ok but as you have experienced are not to be taken for granted.
Far as semiconductors its good practise to double check all ident as one blip difference can make a build become a nighmare.

Anyway you cracked it good work.

Regards