Is Digilent breadboard has high quality?

[NorthGuy]

The wires are longer on solderless breadboards, so the inductance is higher.

I don't think they're too long to cause any significant problem. Also, this is partly compensated by the parasitic capacitance. I worked a lot with breadboards and I've never encountered the problem. At any rate, if you do experience the problem, why can't you simply increase the number or value of bypass capacitors?

[tggzzz]

The wires are longer on solderless breadboards, so the inductance is higher.

I don't think they're too long to cause any significant problem. Also, this is partly compensated by the parasitic capacitance. I worked a lot with breadboards and I've never encountered the problem. At any rate, if you do experience the problem, why can't you simply increase the number or value of bypass capacitors?

That is so wrong on so many levels, I don't know where to begin.

What's the first phenomenon you (should)think of if you have an inductor and a capacitor? What happens when you "hit" it with a step change in voltage or current?

Look, if you can't be bothered to read the references I've given, then you aren't going to learn no matter how much of my time I spend trying to enlighten you.

[NorthGuy]

That is so wrong on so many levels, I don't know where to begin.

Begin with your brain. Try to think a little. Crunch some numbers. Estimate the switching current. Calculate what wire inductance you need to produce significant voltage drops. Calculate how long a wire you need to cause a problem. Calculate how this is all affected by bypass capacitors you may install. Perhaps do some real life measurements to check your numbers. It's not a rocket science, you know. Then perhaps you'll see that you're just making an elephant out of the fly trying to blow the effect out of proportions.

[2N3055]

It's not voltage drops that are your problem. That whole thing is a resonant circuit that rings like crazy.. If you take DIP meter close to it will react. You could design RF oscillator with a FET few resistor and protoboard as a resonant tank..

[tggzzz]

That is so wrong on so many levels, I don't know where to begin.

Begin with your brain. Try to think a little. Crunch some numbers. Estimate the switching current. Calculate what wire inductance you need to produce significant voltage drops. Calculate how long a wire you need to cause a problem. Calculate how this is all affected by bypass capacitors you may install. Perhaps do some real life measurements to check your numbers. It's not a rocket science, you know. Then perhaps you'll see that you're just making an elephant out of the fly trying to blow the effect out of proportions.

Go on then - do it yourself. State the equations and assumptions and plug in the numbers. Prove I'm wrong. Write it up, you will have made your name and people will flock to your door, and you will be able to charge outrageous consulting fees. Seriously.

For bonus points, indicate clearly why people with just a little experience like Winfield Hill (one of the authors of The Art Of Electronics that I referred to) have erroneous calculations and measurements. Or Howard Johnson. Or the device manufacturers.

Hint1: tell us what do you think does happen with an LC circuit?

Hint2: inductors don't always "produce significant voltage drops"

Hint3: here's part of the reference I gave and you have ignored. Note that the PCB conditions are far better than you could find on a solderless breadboard.

[NorthGuy]

Go on then - do it yourself. State the equations and assumptions and plug in the numbers. Prove I'm wrong. Write it up, you will have made your name and people will flock to your door, and you will be able to charge outrageous consulting fees. Seriously.

I did write a blog about signals and breadboards, but you haven't even read it. I guess if I write about ground bounces you won't read it neither

Hint1: tell us what do you think does happen with an LC circuit?

It is quite a bit in my blog on what I think happens in LC circuits on breadboards, with models, diagrams etc. It's not too late. You can click the link and read it.

That's an interesting graph. It is taken from 1988 TI application note ("Advanced  CMOS logic
Designer's Handbook"), where they discuss the pin location DIP packages for their new HC series. They decided to relocate the power pins to the center of the chip. The picture above is the measurements with DIP package having power pins on the sides. When they have put the power pins into the middle, the effect has been alleviated and voltage spike became negligible.

TI certainly wanted to exaggerate the effect to show the advantages of their clever innovative pin placement. They drive 7 50 pF loads. The is equivalent of suddenly connecting 350 pF capacitor to the power rail. Voltage sagged 2V. Today, if this happened to me, I would say that it's either a bad power supply or inadequate bulk/bypass capacitors. Cannot relate to what it would mean in 1988, but anyway, the voltage sagged 2V.

Sane people don't drive capacitive loads directly. It's a good idea to add a series resistor, which limits current and completely eliminates the inrush effect. Thus, the effect of such magnitude cannot be observed in real circuits. This has nothing to do with breadboards, BTW.

Note that the PCB conditions are far better than you could find on a solderless breadboard.

In this particular case, the extra inductance of the signal wires would help. TI had the capacitive loads very close. If there was a foot of wire between the driver and the 50 pF load, none of this would happen. The inductance of the wire would limit the rate of current rise thus eliminating the magnitude of the inrush current. No way you can achieve super fast rise times when driving through long wires. Thus, there wouldn't be any inrush current in the first place.

Not only that. You should expect all the problems caused by too sharp driving to be alleviated by wire inductance in the breadboards. It simply doesn't have enough bandwidth to pass through very sharp pulses. When I was writing my blog, I wanted to find reasonable values of inductance and capacitance which would cause high frequency (500 MHz or so) oscillations, and I failed. I still published the picture of what I wanted to achieve, but the capacitance and inductance I used to model it were out of the range of what can be found in breadboards.

[tggzzz]

You should expect all the problems caused by too sharp driving to be alleviated by wire inductance in the breadboards. It simply doesn't have enough bandwidth to pass through very sharp pulses. When I was writing my blog, I wanted to find reasonable values of inductance and capacitance which would cause high frequency (500 MHz or so) oscillations, and I failed. I still published the picture of what I wanted to achieve, but the capacitance and inductance I used to model it were out of the range of what can be found in breadboards.

Damped inductance in the right place can help, e.g. ferrite beads.
Undamped oscillation rarely helps in digital circuits. In signal wires it causes excess current in clamp diodes, which can upset circuits.
Inductance in the ground lead never helps.

I fail to see what is diffiicult about getting VHF/UHF oscillation. A 6" wire (i.e. 150nH) and 15pF (i.e. 3 gate loads) oscillates at ~100MHz. That's easy to demonstrate and is a problem even with well constructed circuits: all you have to do is use a typical *10 probe, and the probe itself resonates! Hence the benefit of short/zero length ground leads and (expensive) low capacitance probes. Alternatively you could just look at some of the simulations in your blog!

Modelling can only be as useful as the accuracy of the model and the type of the analysis. Your blog, fig 10, is unrepresentative of a solderless breadboard circuit constructed by a beginner:

• a typical R1 "drive" resistance can be as low as 7ohms, not 100ohms
• your R2 is unreasonably low; 10kohms would be less unrepresentative
• there's no ground lead inductance

No beginner would insert extra resistors to increase R1 and decrease R2. Nobody with experience would bother either; they would just construct it properly in the first place.

If you want to model an unrealistically good beginner's circuit, then you should also model the parasitics associated with the extra resistors and their wires required merely because solderless breadboard is being used.

Your blog only indicates indirect measurements at disrete frequencies. Real digital signals have a spectrum that is more complex than that. Such complexity is best modelled anaytically in the frequency domain. That's easy to do with, say, LTSpice. If you can't do frequency domain measurements, then eye-diagrams are an excellent direct measurement technique.

I suggest you re-analyse your model in the frequency domain with a source resistance of 7ohms or so, a receiver resistance of 10kohms, and a ground lead inductance of say 50nH.

[NorthGuy]

No beginner would insert extra resistors to increase R1 and decrease R2. Nobody with experience would bother either; they would just construct it properly in the first place.

I can assure you, the same laws of physics exist on every breadboard in the world, whether it's used by a beginner or by anybody else. It is these laws we're interested in. It is immaterial what someone would or wouldn't do.

If you want to model an unrealistically good beginner's circuit, then you should also model the parasitics associated with the extra resistors and their wires required merely because solderless breadboard is being used.

It wasn't the goal to build an exhaustive model of the breadboard. The goal was merely to test if you can use high speed digital signals on a breadboard.

If someone approaches an engineer and says - look I want to prototype a 100 MHz digital circuit on breadboard, and I need a 2 GSa/s logic analyzer for that - in 90% of the cases the engineer would say that this is a crazy idea, 100 MHz is too fast and 2 GSa/s is a huge overkill. So, I decided to check by myself and the result is quite decisive:

- Yes you can test 100 MHz on breadboards. 200 MHz might be Ok too.
- 2 GSa/s logic analyzer is quite adequate in this situation.

Your blog only indicates indirect measurements at disrete frequencies. Real digital signals have a spectrum that is more complex than that. Such complexity is best modelled anaytically in the frequency domain. That's easy to do with, say, LTSpice. If you can't do frequency domain measurements, then eye-diagrams are an excellent direct measurement technique.

Digital sampling. That's the whole point. You're not really interested in details of frequency content of your digital signals. The only thing that matters is how it's sampled. Be it a perfect square wave or something which more resembles sine wave, or something with huge harmonic distortions, none of this matters if you get the digital sampling correctly.

I suggest you re-analyse your model in the frequency domain ...

My model, albeit simple, gave me the answers I was looking for, and they matched the real experiments well enough. I don't think there's a need for further modeling.

However, if would be interesting to look at real-world eye diagrams, perhaps some frequency domain analyses with a network analyzer might be interesting. It would be especially interesting to analyze failure modes. But, like I said, I don't have adequate scope/probes. If you do, make the experiments. If you do it reasonably well, I'll be happy to publish it.

As to the resistor values - you use what works (which is something not too far from critical damping or slightly over-damped), not what's popular.

[tggzzz]

I've reinserted the relevant context that you snipped for some reason or other, since it is beneficial in understanding why I made that statement.

Modelling can only be as useful as the accuracy of the model and the type of the analysis. Your blog, fig 10, is unrepresentative of a solderless breadboard circuit constructed by a beginner:

• a typical R1 "drive" resistance can be as low as 7ohms, not 100ohms
• your R2 is unreasonably low; 10kohms would be less unrepresentative
• there's no ground lead inductance

No beginner would insert extra resistors to increase R1 and decrease R2. Nobody with experience would bother either; they would just construct it properly in the first place.

I can assure you, the same laws of physics exist on every breadboard in the world, whether it's used by a beginner or by anybody else. It is these laws we're interested in. It is immaterial what someone would or wouldn't do.

I don't think you are adding to the sum of human knowledge by stating "the same laws of physics exist on every breadboard".

This thread is largely about what beginners would and would't do, if you go back and read the thread.

If you want to model an unrealistically good beginner's circuit, then you should also model the parasitics associated with the extra resistors and their wires required merely because solderless breadboard is being used.

It wasn't the goal to build an exhaustive model of the breadboard. The goal was merely to test if you can use high speed digital signals on a breadboard.

If someone approaches an engineer and says - look I want to prototype a 100 MHz digital circuit on breadboard, and I need a 2 GSa/s logic analyzer for that - in 90% of the cases the engineer would say that this is a crazy idea, 100 MHz is too fast and 2 GSa/s is a huge overkill. So, I decided to check by myself and the result is quite decisive:

- Yes you can test 100 MHz on breadboards. 200 MHz might be Ok too.
- 2 GSa/s logic analyzer is quite adequate in this situation.

Unfortunately that isn't correct: the modelling result isn't as decisive as you might like to believe, for the reasons I gave. Basically

• you have only done a test at a few frequencies, not at all the frequencies present in a digital signal
• the tests are't representative of the majority of problems found in these situations

Your blog only indicates indirect measurements at disrete frequencies. Real digital signals have a spectrum that is more complex than that. Such complexity is best modelled anaytically in the frequency domain. That's easy to do with, say, LTSpice. If you can't do frequency domain measurements, then eye-diagrams are an excellent direct measurement technique.

Digital sampling. That's the whole point. You're not really interested in details of frequency content of your digital signals. The only thing that matters is how it's sampled. Be it a perfect square wave or something which more resembles sine wave, or something with huge harmonic distortions, none of this matters if you get the digital sampling correctly.

Firstly, they aren't digital signals, they are analogue signals that are being interpreted by the receiver as digital signals.

Secondly, referring to "harmonic distortions" leads me to doubt you understand the spectrum of digital signals. That is congruent with your not using frequency domain modeling.

More importantly, if the eye is closed due to distortions in the channel, then there is no correct sampling instant!

I suggest you re-analyse your model in the frequency domain ...

My model, albeit simple, gave me the answers I was looking for, and they matched the real experiments well enough. I don't think there's a need for further modeling.

It is easy to create models to give you the answers you seek. Boring.

It is important to use models to give you insight into phenomena that you might need to understand but don't realise are an issue in the first place. In this case, inductance in the ground lead is important.

However, if would be interesting to look at real-world eye diagrams, perhaps some frequency domain analyses with a network analyzer might be interesting. It would be especially interesting to analyze failure modes. But, like I said, I don't have adequate scope/probes. If you do, make the experiments. If you do it reasonably well, I'll be happy to publish it.

Not having the experimental equipment is one reason why modelling is important - provided your model is up to the task (which isn't easy and requires experience).

If an implementation is functioning as designed but doesn't cope with the real world, I don't think "failure modes" is the right term. "Inadequate design" would be closer to the mark.

As to the resistor values - you use what works (which is something not too far from critical damping or slightly over-damped), not what's popular.

You are avoiding the point that people don't put source and termination resistors in solderless breadboard circuits.
Beginners don't realise they are necessary.
Those that understand why they are necessary wouldn't use solderless breadboards in the first place!

[NorthGuy]

You are avoiding the point that people don't put source and termination resistors in solderless breadboard circuits.
Beginners don't realise they are necessary.
Those that understand why they are necessary wouldn't use solderless breadboards in the first place!

This is, how do you call it, "strawman argument". I made no claims on how people use breadboards, I am not interested in the usage patterns at all. I'm arguing about the physical things happening in breadboard circuits. Resistors are irrelevant here. 100 MHz signals worked without resistors in all cases. 200 MHz signals may require resistors. If you were using PCB, I'm sure you would terminate your signals as needed. Why is it wrong to use termination on breadboards?

My point is very simple. I did some modeling which suggested that high speed signals would work, then I run the tests, and they worked. The bottom line is that if I need to run some tests with 100 MHz digital signals, I may consider doing this on a breadboard, which I would never do before. That's a benefit to me, and perhaps to other people who may decide likewise. End of the story.

[ebastler]

End of the story.

I sure hope so. Alas, I lack the faith... 

[Oak]

Digilent breadboard is product of known Wisher, I asked via mail and they answered yes.

Hello,

I wanted to buy a WBU-206 breadboard but they said that it isn't in stock and they also don't know when it will be available for sale. They recommended Digilent breadboard, they said it's a product of Wisher and same quality as yours. I just want to know if it's your product or not.The company is your Turkey distributor Elektrovadi.

Thank you very much in advance.
Oguz Akilli

PRODUCT LINK:
https://www.elektrovadi.com/urun/solderless-breadboard-kit-large

They answered:

Wisher Enterprise Co., Ltd <wishmake@ms59.hinet.net>
Wed, 13 Oct, 09:04
to me, wishmake

Dear Oguz Akilli

 

Thank for your trust and support to WISHER.

 

We are OEM manufacurer from Digilent,

 

this item no. 3400002-1 of Digilent is manufactured by us certainly.

 

The quality of this item from Digilent is same as Wisher.

 

If you are willing to purchase our product, you may buy it from them directly.

 

If you have any question or concerns , please feel free to contact me.

 

I wish you all the best in your future, thank you!

 

================================
Regards,

Eva

Sales Representative

WisherLogo1

WISHER ENTERPRISE CO., LTD. 
4F-2, No.333, Fu-Hsing N. Rd.,

Taipei City 10544, Taiwan R.O.C.

Tel:+886-2-2719-1369
Fax:+886-2-2719-1366

Web: www.wishmaker.com.tw