So after being here for a while ::) I have decided to try designing a delay circuit from some of the Reticon R510x chips that I have. I am using the datasheet circuit as a starting point however, I have neither the design experience nor the complete understanding of different audio "stages" to move forward.
The datasheet I am referring to is here:
http://www.experimentalistsanonymous.com/diy/Datasheets/RD5108.pdf
Specifically Sheet 4. As you can see, the input is shown DIRECTLY into the BBDs input via an electro and being biased with the circuit's Vdd voltage. So my question is...
Doesn't there need to be some form of input buffering or an input gain stage via an op amp or transistor network? Any other delay based circuit I have worked on has NEVER gone directly into the BBD. Could someone perhaps point me to some literature that will help me understand this particular project?
From what I see, it appears that Q2 and the 741 are in place to create some form of recovery stage but, I am probably talking out of my arse as I have no clue :icon_redface:
Please help me understand what is going on here. Thanks! ;D
Go for it!
This may be a silly suggestion from a novice, but maybe just breadboard it and see? Play around with it. If you think it needs more oomph in front, copy an input stage from something simple, play around with the gain in front of it?
It would be awesome if you kept a design log/diary kinda thing to share afterwards (even if it was very rough)
George
Thanks George!
I realize this is an extinct chip BUT, I have a few and would like to use them for SOMETHING!! 8)
This is new for me as I have not gotten into anything this deep on a "design" level. I am usually just following directions and/or making tiny, tiny leaps into the unknown.
I am hoping to get some form of input however, from the research I have done, it looks like there are not many that like to deal with BBD based circuits.
Not looking for a handout persay... Just looking for some direction ;D
The datasheet indicates a typical input of 1.5V RMS....therefore yes, unless you want your raw guitar signal to be lost in the IC's noise floor, then you'll need a gain stage in front.
The input 'shunt resistance' is 300k (which if I'm interpreting that correct is probably the IC's input impedance....not outrageously low, but a gain stage would eradicate any worries on that aspect.)
So, I'd say look at putting an an adjustable gain stage in front of the IC ....1,5V RMS is about 4V peak to peak....a typical guitar signal initial transient is about 1V peak to peak (though this can vary with pickups & playing style), so I'd say a gain stage adjustable between 3x and 8x will get you in the ballpark.
Depending on your follow on circuity you might want to take some of the magnitude out of the output signal too (a simple output pot) as 1.5V RMS may well clip any follow on stages.
Quote from: Gurner on February 17, 2013, 07:50:30 PM
The datasheet indicates a typical input of 1.5V RMS....therefore yes, unless you want your guitar signal to be lost in the IC's noise floor, then you'll need a gain stage.
The input 'shunt resistance' is 300k (which if I'm interpreting that correct is probably the IC's input impedance....not outrageously low, but a gain stage would eradicate any worries on that aspect.
SO I'd say look at putting an an adjustable gain stage in frontvof the IC ....1,5V RMS is about 4V peak to peak....a typical guitar signal initial transient is about 1V peak to peak (though this can vary with pickups & playing style), so I'd say a gain stage adjustable between 3x and 8x will get you in the ballpark.
I appreciate your input Gurner... Thanks ;)
Unfortunately, I have no idea how to construct or incorporate said gain stage :icon_redface: Could you point me in the right direction?
It's a little small, but this is the kind of circuit you'd want....
http://i.cmpnet.com/rfdesignline/2007/06/TI_pt1_Fig2.gif
google non inverting opamp stage .....heaps of explanations better than I could muster!
Thanks again Gurner!
I see lots of examples for a non-inverting gain stage BUT, they are all simple compared to the example that you show above. Just the op amp and 2 resistors controlling the gain amount. Your example shows several resistors and some caps which I know what a few will do but, mostly I have no idea :icon_redface:
I feel such the NOOB!! :-\
After breifly looking at the data sheet and seeing that the 5108 is a positive voltage BBD with voltage ratings of 7 to 13 volts, I don't see why this delay IC can't be retrofitted into a MN3208 or MN3205 circuit.
You can at least use those circuits as a guide to work up your own build.
The circuit on page 4 is an evaluation circuit not a production delay unit.
Quote from: armdnrdy on February 17, 2013, 09:12:49 PM
The circuit on page 4 is an evaluation circuit not a production delay unit.
I understand this. I was merely using it as a place to start. Although, I do not see why it could not be implemented. Unless I am missing something :-\
Also, this is only a 1024-stage delay. The MN3208 and MN3205 are 2048 and 4096 respectively. I am looking at this as something new, not something I want to re-invent. Want to learn some stuff along the way as well ;)
Gotcha!
Your first post threw me when you stated that you wanted to design a "delay circuit" with this IC.
Well then I would look to some MN3207 circuits for guidance if you don't know where to start. That's what Scott Swartz did with the AD-3208. He pulled a little from here and a little from there.
@Larry
Ultimately, I would like to design a delay circuit based on this chip. I would just like it to be unique and I would like to go through the process as I have not done that before from start to finish. I figured the evaluation circuit was a good place to start and I could build off of it. There are just a lot of things I do not understand and need to go through the process for the knowledge.
Might stop by here constantly for a point in the right direction though ;D
This is the best I have read on op-amps and using them to amplify a signal, or buffer, mix, filter:
http://sound.westhost.com/dwopa.htm
If you really need to cut to the chase, scroll down to inverting opamps and noninverting op-amps. However, if you can wrap your head around the third heading "Basic Rules of Opamps" you will never regret it!
(http://i.cmpnet.com/rfdesignline/2007/06/TI_pt1_Fig2.gif)
My brief description of what the parts in this circuit do (Left to right)
Rt is an input resistor (so good nice and high)
1. First cap is an input cap. This has to do a few things:
- Allow all the frequencies that we need through (audible guitar frequencies)
- Stop any DC voltage that might be coming into the circuit from the input
- Contain the DC voltage that we're adding at the next stage which is...
2. The next unlabelled resistors from V+ and from ground set the bias point by way of a voltage divider. Use the voltage divider equation to work out what the added DC voltage is to the signal path by Bias voltage = V+ * {Rbottom / (Rtop + Rbottom)} This is the bias voltage needed by the particular op amp you use
3. The next two resistors are what define the gain of the op amp:
(http://www.elexp.com/tips/noninvrt.gif)
4. This whole section of the circuit is only dealing with the small AC signal, and we contain this part of the circuit with the cap at the bottom and the output cap (only the output cap needs to be selected to allow guitar frequencies through). Therefore it's all effectively working with the signal voltage but shifted up by the DC voltage set by stage 2
5. The final resistor is the output resistor (good low!) and there is an example of a load on the far right.
Hope this helps. Even with knowing the function of each part, it's probably easiest to mirror another circuit to save time on calculating each part. Then you know how to play with individual areas to improve the performance if you know what each part is meant to do.
Good luck!
George
(I hope my electronics modules are serving me well here, please correct anyone who spots inaccuracies)
Thanks Ian and George!!!
Great stuff and great explanations. Gonna soak this in and try experimenting! ;D
Sounds like a fun project to get into. Ian's link is very good - read that one before. Good ol' Wikipedia has a good introduction to opamp circuits (here (http://en.wikipedia.org/wiki/Operational_amplifier_applications)) and some useful links at the bottom of the page ("Op Amps for Everyone" is a perennial favourite). I guess you're going to need buffers and mixers in there, and opamps are a neat and easy way to do these. Or some simple one-transistor circuits (e.g. at AMZ).
Looking forward to seeing how this pans out. Good luck!
Just to add to Geroge's description, the two final resistors (far right of schem) will not be needed in your situation (they assume you don't know what follows on...but in your situation you do & they won't be needed)
Additionally, if you set the opamp input bias point to about half the supply voltage (those two resistors attached to the +ve input ....use identical values ...about 1M or higher), then the output cap won't be needed either...as the DC level on the ouput of the opamp will be half the supply voltage....which is what your IC will require (ish)
Right now, I am thinking about setting the Vbias resistors at 1M each. Also, I am going to start with an R2 value of 100K and install a 1M trimmer for R1 (initially set at about 300K) and tune until I get a good input to the BBD.
Can I use the 1uF electro as the op amp's output cap? (The one that is on the input to the BBD)
Also, is there going to be a problem with the input gain stage bias VS. the BBD bias? I can assume that running at 12V, the input gain bias would be ~6V. I am wondering if that will be a problem for the BBD input bias point ???
Quote from: Govmnt_Lacky on February 18, 2013, 11:07:32 AM
Right now, I am thinking about setting the Vbias resistors at 1M each. Also, I am going to start with an R2 value of 100K and install a 1M trimmer for R1 (initially set at about 300K) and tune until I get a good input to the BBD.
Can I use the 1uF electro as the op amp's output cap? (The one that is on the input to the BBD)
Also, is there going to be a problem with the input gain stage bias VS. the BBD bias? I can assume that running at 12V, the input gain bias would be ~6V. I am wondering if that will be a problem for the BBD input bias point ???
As I said in my post immediately above yours, if you arrange the DC bias for opamp +ve input pin right, then you also satisfy the bias required for the follow on IC... your IC's datasheet says the input bias should 0.6 VDD (Table 2 ...'input signal bias'), therefore if you arrange the opamp bias resistors with a 40/60 ratio, then that'll do nicely (eg top resistor 800k, bottom resistor 1.2M, then your DC level on the output of your opamp will be the ideal bias for your follow on IC & you won't need an electro output DC blocking cap.
@Gurner
Ok. Just to clarify...
If I set the resistor between V+ and the op amps non inverting input at ~800K and set The resistor between the op amps non inverting input and GND at ~1.2M then I should have the input gain stage biased just about right to send the signal directly from the input op amp's output directly to the BBDs input without the need for the 1uF DC blocking cap?
You also mentioned above that I should aim for a 4X to 8X output from the op amp gain stage. Does that still hold water with this basing arrangement? I would think so ;D
Cheers and thanks for helping me understand this!!!
Yepp that will put your bias at around 7.2V so that 1uF blocking cap isn't needed.
Then I'd use a variable resistor for R1 and a set resistor R2 which is about a tenth of R1's max value. This will let you play with the voltage gain from unity gain to about 10X gain.
(R1 and R2 are based on the second diagram in my last post, the labels are too small for me to read in the first diagram)
This gain amplifies the varying voltage (i.e our signal) but the signal stays shifted up at 7.2V above ground. Roughly put, its like a sinewave with an average voltage of 7.2V not 0V. It doesn't get shifted up or down on the graph, just gets a bigger sine wave amplitude.
Have fun!
George
Quote from: Govmnt_Lacky on February 18, 2013, 01:00:08 PM
@Gurner
Ok. Just to clarify...
If I set the resistor between V+ and the op amps non inverting input at ~800K and set The resistor between the op amps non inverting input and GND at ~1.2M then I should have the input gain stage biased just about right to send the signal directly from the input op amp's output directly to the BBDs input without the need for the 1uF DC blocking cap?
yes that's right, if I'm reading your IC's datasheet correct! (the datasheet say the input to the IC should be set at about 0.6 of VDD)
Quote from: Govmnt_Lacky on February 18, 2013, 01:00:08 PM
You also mentioned above that I should aim for a 4X to 8X output from the op amp gain stage. Does that still hold water with this basing arrangement? I would think so ;D
Yes it does, don't get bias confused with gain, regardless of the bias arrangements the gain needed works out something like....
'IC's max input level' divided by your 'guitar's maximum output signal'
To get the best signal to noise ratio, you want sufficient gain to get the maximum signal through your IC as possible (but no too much gain that your IC input pin clips)
....for example, if the IC is expecting 4V peak to peak
max & you guitar outputs 1V peak to peak
max, then you'd need the opamp to have a gain of 4 or less (a bit less actually to allow a little headroom & avoid clipping) .....this can never be an exact science, because there are many variables to do with your guitar ouput - string gauge, string material, pickup type, guitar playing style etc...this is why I recommended having a variable gain trimpot to tweak your opamp's gain to suit.
Quote from: Gurner on February 18, 2013, 01:21:23 PM
...this is why I recommended having a variable gain trimpot to tweak your opamp's gain to suit.
That is the plan! ;D
A 100K from the inverting input to ground AND a 1M trimmer from the inverting input to the op amp output. This way, I can tweak the trimmer until I get the correct gain.
At least, that is what I interpret is a good way to adjust the output gain.
-OR-
Would it be better to have a set value between the inverting input and op amp output, then... put a trimmer between the inverting input and GND?
One last question before I get this on the breadboard..
Is there a reason why I need to use "higher" values for the biasing resistors?
Instead of using the 1.2M/800K network to get the bias, could I use smaller values like 470K/330K? It is not EXACTLY 40/60 but it is pretty close.
Both choices you've suggested draw currents in the range of uA (I believe 6uA and 15uA) rather than mA which is what we're after. We obviously avoid 100 ohm + 100 ohm because that would be drawing 60mA, which would be double what the whole pedal uses.
In terms of power there's no difference really. I might have missed something in a different aspect, but I can't really see what else would be relevant. Go for it!
George
(PS with resistor tolerances, it probably doesn't make a huge difference with the bias point precision, but theoretically 470K and 330K are more accurate to 7V)
Quote from: gcme93 on February 18, 2013, 03:46:01 PM
Both choices you've suggested draw currents in the range of uA (I believe 6uA and 15uA) rather than mA which is what we're after. We obviously avoid 100 ohm + 100 ohm because that would be drawing 60mA, which would be double what the whole pedal uses.
In terms of power there's no difference really. I might have missed something in a different aspect, but I can't really see what else would be relevant. Go for it!
George
(PS with resistor tolerances, it probably doesn't make a huge difference with the bias point precision, but theoretically 470K and 330K are more accurate to 7V)
Some of this is above my head :icon_redface:So are you saying that the 800K/1.2M Vref biasing network is TOO big? Do you have a better suggestion for values?After reading this again, I see what you are meaning! :icon_rolleyes:
I also see why I need to get the output voltage of the op amp gain stage to 7.2V as this is the required input to the BBD (0.6 of 12V which is Vdd)
So essentially, keeping the values higher for the Vref resistor network will also reduce the overall current draw of the circuit... right?
If I am correct.... then why don't ALL circuits that require an op amp gain stage use higher value resistors for biasing in order to keep the current lower? Is it an availability issue in manufacturing?
*Post modified to answer new question as you understand the old point ;) *
Yepp, high values means less current is drawn.
In the scheme of things, your values don't need to go that high because the difference in current being drawn is almost nothing compared to the rest of the circuit.
If we used a ratio of 1.2K to 0.8K, the current would be 6mA which is a little too much.
However, just one multiple of ten up (12K and 8K) and the current drawn is now 0.6mA which is pretty darn low. To make it a pretty much negligible amount, 120K and 80K draws 0.06mA
I guess you're getting the point now. We want the biasing resistors to add up to about 500K or more to make sure the currents they draw are tiny. Past that all that matters is getting a suitable ratio for setting the voltage right :)
(in terms of manufacturers, maybe it's cheaper to use lower resistance parts? Maybe there's another tiny influence on performance?)
I just want to say a great big...
THANK YA VERY MUCH!!!
to all who have helped me understand what the heck I am looking at!!
Without the shared knowledge on this site it would take me 10x longer to understand!!!
More to come.... HOPEFULLY ;)
Re the high value input resistors, it's not just to do with aiming for less current draw, but if you're plugging in a raw guitar signal, you need to keep the input impedance of your circuit high (else you'll experience treble rolloff on your guitar's signal).....that's why it's best to stick with resistances in the order of 1M.
Quote from: Gurner on February 18, 2013, 05:19:31 PM
Re the high value input resistors, it's not just to do with aiming for less current draw, but if you're plugging in a raw guitar signal, you need to keep the input impedance of your circuit high (else you'll experience treble rolloff on your guitar's signal).....that's why it's best to stick with resistances in the order of 1M.
+1
Ahh of course! I knew I was forgetting something ;)
Quote from: Gurner on February 18, 2013, 05:19:31 PM
Re the high value input resistors, it's not just to do with aiming for less current draw, but if you're plugging in a raw guitar signal, you need to keep the input impedance of your circuit high (else you'll experience treble rolloff on your guitar's signal).....that's why it's best to stick with resistances in the order of 1M.
Are you speaking of the Vref resistor network -OR- the resistors that set the op amp's gain?
I would like to keep the current draw down as not to limit my powering options. I will start with the 100K/1M trimmer for the gain and go from there.
I dont mind using the 800K/1.2M Vref network however, I just wish I could find a more common value than 800K :-\ I guess I will make due.
Quote from: Govmnt_Lacky on February 18, 2013, 05:53:44 PM
Are you speaking of the Vref resistor network -OR- the resistors that set the op amp's gain?
I dont mind using the 800K/1.2M Vref network however, I just wish I could find a more common value than 800K :-\ I guess I will make due.
It's the vref resistor network I was referring to....don't get too hung up on the exact values....for example, in the earlier random example I gave (on a whim, I probably picked hard to source resistor values!) 750k top resistor & 1.2M lower resistor should be close enough ......you are ideally shooting for a 40:60 ratio (top resistor:bottom resistor in your vref resistor chain) & where the resistor values are common & easily sourcable ....therefore if you use a 1M lower resistor, the upper resistor needs to be (1,000,000/60) * 40 = 666k or if you use a lower resistor of 1.5M then the top resistor needs to be (1,5000,000/60) * 40 = 1M
UPDATE:
Got this on the breadboard and it looks like I will need to tear it all down and start again :-\
After applying power to take some measurements, my regulated PS was showing a current draw of about 600mA :icon_eek:
It slowly crept down to 200mA and I could detect the ever so slight odor of burning failure! :icon_evil: I believe it was coming from the 4013 clock chip.
I am thinking that I may have hooked something up incorrectly OR some of the components were touching and they were not supposed to. I will need to re-build it and give myself a bit more space.
More to follow ;D
There is a quick check for breadboards drawing to much current - before applying power, measure the resistance between power (+9V) and ground. If you have say 100 ohms resistance, then the current will be at least V=IR, 9=I x 100, I= 90mA. Most chips draw a few mA each, bias resistors much less, its a good guess that most analogue circuits will draw less than say 30mA. So if you measure less than 300 ohms, then double check your circuit. Most circuits I have built measure well over 1K from power to ground. After I cooked a battery, I always do this check.
Quote from: Ronan on February 26, 2013, 06:19:18 AM
There is a quick check for breadboards drawing to much current - before applying power, measure the resistance between power (+9V) and ground. If you have say 100 ohms resistance, then the current will be at least V=IR, 9=I x 100, I= 90mA. Most chips draw a few mA each, bias resistors much less, its a good guess that most analogue circuits will draw less than say 30mA. So if you measure less than 300 ohms, then double check your circuit. Most circuits I have built measure well over 1K from power to ground. After I cooked a battery, I always do this check.
Shit... just printed that, and pinned it to the wall over my bench. Thanks for the tip Ian. That's about as KISS as KISS can be.
I forgot to mention that it may take a minute or two for the reading to settle, as the filter cap/s charge up.
To be honest, I have put this on the back burner for now :-\
Between XP ALLs, Ludwigs, Electric Mistresses, and a few more... I have to clear some serious room before "experimenting!" :icon_eek: